Talmy lexicalizationpatterns

---

2 Lexicalization patterns: semantic
structure in lexical forms
LEONAKD TALMY

This chapter addresses the systematic relations in language between
meaning and surface expression. Our approach to this has severa1
aspects. First, we assume we can isolate elements scparately within the
domain of meaning and within the domain of surface expression. These
are semantic elements like ‘Motion’, ‘Path’, ‘Figure’, ‘Ground’, ‘Man-
ner’, anrl ‘Cause’, and surface elements like ‘verb’, ‘adposition’, ‘sub-
ordinate clausc’, and what we will cliaracterize as ‘satellite’. Second, we
examine which semantic elements are expressed by which surface
elements. This relatinnship is largely not orle-to-one. A coml)ination of
semnntic elements can be expressed by a single surface elemcnt, or ü
single semantic element hy a combinatiun of surface elements. Or again,
semantic elemeuts of different types can be exptessed by the same type
of surface elenient, as well as the same type by severa1 dif-t‘erent ones.
We finrl here a range of typological palterns anti universal principies.
We do n»t look at every case of semantic-to-surface association, but
only at enes that constitute a pervasive pattern, either within a language
01’ across lunguages. Our particular concern is to understand how sucll
patterns compare across languages. ‘I’llilt is, Eor a particular semantic
domain, we ask if languages exhibit a wicle variety of patterns, ä
compüratively sniall n~~uiller of patlerns (a typology), or a single pattern
(a universal). We will be intcrested primarily in the last two cases, as
well as in the case where a pattern appears in no languages (universal
exclusioii). Our approach can be Srll~rl~i~~iZetl a s iii tliis procedural
otltliw:

(1) (‘cntities’ = elements, relntions, and striictures: both particular
cases ant1 categories of these)

a. Determine varinus semantic entities in a language.
11. LSelermirle various surlace entities in the language. ’

58 LEONARD TALMY Lexicalization patrerns 59

c. Observe which (a) entities are expressed by which (b) they embody. The conclusion in section 3 compares the advantages of
entities - in what combinations and with what interrelations - the approach adopted here and extends this to the issue of informational
noting any pattems. foregrounding and backgrounding. And, linally, the Appendix in sec-
d. Compare flndings of this sort across languages, noting any tion 4 tabulates and augments the meaning-form relations described in
patterns . the text.3

This outline sketches the broad project of exploring meaninwurface
relations. But our present undertaking is narrower in severa1 ways. We outline now some general characteristics of lexicalization, as part of
First, there are two directions for exploring meaning-surface relations, this chapter’s theoretical context. A meaning can be considered associ-
both of them fruitful. One direction is to hold a particular semantic ated with surface forms mainly by three processes: lexicalization,
entity constant and observe the surface entities in which it can appear. deletion (or zero), and interpretation. We can contrast these three in an
For example, one could observe that the semantic element ‘negative’ example where no one proc&s clearly applies best. Consider the phrase
shows up in English as a verb-complex adverb (will nof go), as an what pressure (as in What pressure wus exerted?), which asks ‘what
adjective (no money), as an adjectival derivational af&x (unkind), and degree ofpressure’ - unlike the more usual what color, which asks for a
as a verbal incorporated feature (doubf); in Atsugewi as a verb requiring particular identity among altematives. We could account for the ‘de-
an infinitive complement (miPi:p ‘to not’); and in some languages as a gree’ meaning by lexicalization: pressure here differs from the usual
verbal inflection. The other direction is to hold constant a selected usage by incorporating an additional meaning component: pressure2 =
surface entity, and to observe which semantic entities are variously &gree of pressure, (or, altematively, there is a special whar here:
expressed in. it. The present chapter explores in only this second whaf, = whaf, degree 08. Or we could assume that some cunstituent
direction. like degree of has been deleted from the middle of the phrase (or
Within this limitation, we narrow our concerns still further. One can that a zero form with the meaning ‘degree of now resides there). Or
examine surface entities of different morpheme count for the meanings else, we could rely on a process of semantic interpretation, based on
that appear in them. At the small end of the scale are the ‘zero’ forms. present context and general knowledge, to provide us with the ‘degree’
Thus, by one interpretation, there is a missing verbal expression in meaning.4
English constmctions like I feel like [having] a milk shake and I hope for In general, we assume here that Iexicalization is involved where a
[fhere fo be] peace, or in German ones like Wo wollen Sie denn hin particular meaning component is found to be in regular association with
[gehen/fahren/ . . .]? ‘Where do you want to go?‘. One might conclude a particuIar morpheme. The study of lexicalization, however, must also
that such missing verbal mean@ come from a smaI1 set, with members include the case where a set of meaning components, bearing particular
like ‘have’, ‘be’, and rgo’.2 Altematively, one could investigate the relations to each other, is in association with a morpheme, making up
meanings that are expressed by surface complexes. A comparatively the whole of the morpheme’s meaning. In the clearest case, one
lengthy construction might encode a single semantic element. Consider morpheme’s semantic makeup is equivalent to that of a set of other
the approximate semantic equivalence of the construction be of inreresr morphemes in a syntactic construction, where each of these has one of
to and the simple verb interest, or of carry out an investigation into and the original morpheme’s meaning components. A familiar example here
investigare. However, this chapter Iooks only at the middle size Ievels - is the appioximate semantic equivalence between kill and make die.
single morphemes and, to a lesser extent, words (composed of root and However, such clear cases are only occasional: it would be unwise to
derivational morphemes) . base an approach to lexicalization on semantic equivalences solely
In particular, we will investigate one type of open-class element, the between exfant morphemes. What if English had no word die? We
verb root,.the topic of section 1, and one type of closed-class element, would still want to be able to say that kiil incorporates the meaning
the ‘satellite’, defined and treated in section 2. These two surface types component ‘cause’ (as we would for the verb (to) poison ‘kill by poison’,
are vehicles for a connected set of semantic categories. Our aim in these which in fact lacks a non-causative counterpart for ‘die by poison’). To
sections is to set forth a class of substantial meaning-in-fop language this end, we can establish a new notion, that of a morpheme’s muge: a
pattems, and to describe the typological and universal principles that particular selection of its semantic and syntactic properties. We can then

60 LEONARD TALMY Lexicalization pattems 6 1

point to usage equivalences between morphemes, even ones with motion event consists of one object (the ‘Figure’) moving or located
different core meanings, and even across different languages. To with respect to another object (the referente-object or ‘Ground’). It is
consider one example, there is a usage equivalence between kiil and analyzed as having four components: besides ‘Figure’ and ‘Ground’,
muke appear. Kill indudes in its meaning the notion ‘Agent acting on there are ‘Path’ and ‘Motion’. The ‘Path’ (with a capital P) is the course
Patient’ (‘causative’) and, syntactically, takes an Agent subject and followed or site occupied by the Figure object with respect to the
Patient object; this usage is equivalent to that of make, which incor- Ground object. ‘Motion’ (with a capital M) refers to the presente per se
porates the notion ‘Agent-to-Patient relation’, in construction with in the event of motion or location (only these two motion states are
appear, which incorporates the notion ‘Patient acting alone’ (‘non- structurally distinguished by language). We will represent motion by the
causative’) and takes a Patient subject. Such relationships can be form ‘move’ and location by ‘beL’ (a mnemonic for ‘be located’).’ In
represented, for cases involving both lexical (L) sud grammatical (o) addition to these interna1 components a Motion event can have a
morphemes , as: ‘Manner’ or a ‘Cause’, which we analyze as constituting a distinct
(2) usage of = usage of externa1 event. All these semantic entities can be seen in the following
t? ~~ in construction with G sentences:
(e.g. Ll= kiil, Ll = appear, and G = make)
(4) Manner: Cause:
We can say here that Ll incorporates the meaning of G and that Ll either motion: The pencil rolled off The pencil blew off the
does not incorporate it or incorporates a meaning complementar-y to it. the table table
In the special case where a single morpheme can function equally as Ll location .* The pencil lay on the The pencil stuck on (to) the
or Ll, we can say that it has a range of usages. For example, there is a table table (after 1 glued it)
usage equivalence between break2 and make breakI, as seen in I broke
the vase and I made the vase break, so that break can be said to have a In al1 four sentences, the pencii functions as the Figure and the table as *
usage-range covering both the causative and the non-causative. An the Ground. Ofi and on express Paths (respectively, a path and a site).
equivalent way of characterizing such a usage-range is as in (3). As an The verbs in the top sentences express motion, while those in the
example of this, the causative/non-causative usage-range of break bottom ones express location. In addition to these states of Motion, a
equals the causative usage of kill plus the non-causative usage of appear. Manner is expressed in rolled and lay, while a Cause is expressed in blew
and sruck.
(3) usage-range o f = usage of + usage of The terms ‘Figure’ and ‘Ground’ are taken from Gestalt psychology
L3 L2 Ll but we give them a distinct semantic interpretation here: the Figure is a
where b and Ll are related as in (2) _ moving or conceptually movable object whose path or site is at issue; the
One terminological note: we will refer to the meaning-in-form Ground is a referente-frame, or a referente-point stationary within a
relation with three terms. They are ‘lexicalization’ from McCawley (e.g. referente-frame, with respect to which the Figure’s path or site is
1968); ‘incorporation’ as used by Gruber (1965); and ‘conflation’, a term characterized.6
that was coined for this purpose by the author (Talmy 1972) and that has 1.0 Th2 ved3
now gained general currency. These terms have different emphases and
connotations that will become clear as they are used below, but all refer In this study of the verb, we look mainly at the verb root alone. This is
to the representation of meanings in surface forms. because the main concem here is with the kinds of Iexicalization that
involve a single morpheme, and because in this way we are able to
0.2 Skztch of a motin event compare lexicalization patterns across languages with very different
A number of the pattems looked at below are par-t of a single larger word structure. For example, the verb root in Chinese generally stands
system for the expression of motion and location. We will here provide a alone as an entire Word, whereas in Atsugewi it is surrounded by many
sketch of this system. A fuller analysis appears in Tahny (1975). afflxes that all together make up a polysynthetic verba1 Word. But these
To begin with, we treat a situation containing moveme.nt or the two languages are on a par with respect to their verb roots.
maintenance of a stationary location alike as a ‘motion event’. The basic Presented first are three Iexicalization types for verb roots that

62 LEONARD TALMY . Lexicalization patterns 6 3

together constitute an apgarently exhaustive typology. Any language g . 1 twUred/popped the cork out of the bottle
uses only one of these types for the verb in its most characteristic
self-agentive
expression of Motion. Here, ‘characteristic’ means that: (i) It is
culloquial in style, rather than Iiterary, stilted, etc. (ii) It is frequent in h . 1 ran/limped/jumped/stumbled/rushed/groped m y w a y
occurrence in speech, rather than only occasional. (iii) It is pervasiue, down the stairs
rather than limited, that is, a wide range of semantic notions are i. She wore a green dress to the party
expressed in this type. move + Cause
non-agentive
I .I Motim + Manner/Cause
j. The napkin blew off the table
In a motion-sentence pattem characteristic of one group of Ianguages,
the verb expresses at once both the fact of Motion and either its manner k. The bone pulled Ioose from its socket
or its cause. A language of this type has a whole series of verbs in agentive
common use that express Jnotion occurring in various manners or by 1. 1 pushed/threw/kicked the keg into the storeroom
various causes. There may aIso be a series of verbs expressing lucnfion
m. 1 blew/$icked the ant off my plate
with various manners or causes, but they are apparently always much
fewer. The semantic-to-surface relationship here can be represented as n. 1 chopped/sawed the tree down to the ground at the base
follows: o. 1 knocked/pounded/hammered the nail into the board with
a mallet’
Figure. Motion Fath Ground
{Jy), E:yl To a speaker of a language like English, such sentences may seem so
straightforward that they offer little to ponder. How else might such
propositions be colloquially expressed? But in fact there are languages
with very different pattems of expression. Even a language as seemingly
mrface vertm kindred as Spanish can express virtuaily none of the above sentences in
Figure 2.1 Manner or Cause conflated in the Motion verb the way that English does, as is demonstrated below.
Some indication can be given of the type of conflation pattem
Language families that seem to be of this type are Chinese and involved here by paraphrases that represent the separate semantic
apparently al1 branches of Indo-European except (post-Latin) Ro- components individually - i.e., that ‘unpack’ the sentences. The Manner
mance. English is a perfect example of the type: or Cause notions conflated in the verbs are then best represented by
English expressions of Motion with conflated Manner or Cause separate subordinate clauses, as in the following:
beL + Manner Unconflated paraphrases of English Motion expressions
(6)
a. The lamp stuod/luy/leaned on the table be‘ + Manner
b. The rope hung across the canyon from two hooks. a’. (The lamp lay on the table =)
move + Manner The lamp was-located on the table, iying there
non-agentive b’. (The rope hung across the canyon from two hooks=)
c . The rock sIid/rolled/bounced down the hill The rope was-located [extended] across the canyon, hang-
ing from two hooks
d. The gate swunglcreaked shut on its rusty hinges
move + Manner
e. The smoke swirled/squeezed through the opening
non-agentive
agentive
c’. (The rock rolled down the hill =)
f . 1 slid/rolled/bounced the keg into the storero& The rock moved down the hill, rolling [the while]

64 LEONARD TALMY Lexicalization p a t t e m s 65

d’ . (The gate swung shut on its rusty hinges =) We can represent the relationship between the two meanings offloat in
The gate moved shut [= shut], swinging on its rusty hinges this way:
[the while]
(IO) The craft moved [floating, (the while)] into the hangar on a
agentive cushion of air
floated2
f’. (1 bounced the keg into the storeroom =)
1 moved the keg into the storeroom, bouncing it [the while]
self-agentive The final occurrence here of phrases of two different kinds - the
directional into the hangar and the locative on a cushion ofair - support
h’ . (1 ran down the stairs =) the interpretation that this verb conflates two otherwise separate
I went down the stairs, running [the while] concepts, one of motion and one of locative relationship: each com-
i’. (She wore a green dress to the party =) ponent of the verb is, at least semantically, ‘in censtruction with’ a
She went to the party, wearing a green dress [the while] different one of the two final phrases.
move + Cause The same pair of usages can be seen in an agentive verb such as kick.
non-agentive In its basic usage, this verb refers to an agent’s impacting his/her foot
into some object, but entails nothing about that object’s moving. This is
j’ . (The napkin blew ofE the table =)
obvious when that object is understood in fact to be fured in place:
The napkin moved off the table, from [the wind] blowing on
it (11) 1 kicked, the wall with my left foot
agentive
Again, this verb can be used in a subordinate clause alongside arr
1’. (1 kicked the keg into the storeroom =) independent referente to motion, as in (Iza). Again, it has a second
1 moved the keg into the storeroom, by kicking it usage that incorporates this referente to motion, as in (rzb). And again,
n’. (1 chopped the tree down to the ground at the base =) this latter two-in-one form can link up with a corresponding pair of final
1 moved the tree down to the ground, by chopping on it at phrases, also seen in (Iab) :
the basca
(12) a. 1 moved the ball across the field, by kickingr it with my Ieft
Paraphrase pairs like these reveal a further fact about English: it has a foot
system of lexicaiization doublets. In many cases, a single verb form can
b. 1 kicked, the bah across the field with my left foot
be used either with or without an inccrporated idea of motion. For
example, in its basic usage the verbfloat refers to the buoyancy relation The relation between the two usages here, corresponding to that
between arr object and a medium, and in this sense it is equivalent to be shown in (ro), can be represented as: kick, = ,move [by kickingr] -
afloat, as in: where the subscript ‘*’ indicates the agentive (i.e., ‘cause to move’).g
(7) The craft íIoated/was afloat on a cushion of air We can further support the idea that these verbs each represent two
distinct lexicalizations by showing verbs that have only the one or the
Given the subscript ‘1’ to mark this usage, the verb can aIso appear in a other. Lie as in The pen lay on the plank is semantically much like floarl
subordinate dause, next to a main clause referring to motion: in referring to the support relation between one object and another
(8) The craft moved into the hangar, floatingr on a cushion of air (rather than buoyancy the relationship here is one of linear object in
But the same verb form has a second usage that includes the idea of contact along its length with a firm surface). But it cannot also be used iu
motion together with that of buoyancy. The verb in this usage - here a motion-incorporating sense like Joa& * The pen lay down the incline -
i.e., moved down the incline while in lengthwise contact with it.
marked with the subscript ‘2’ - can appear in a ene-clause sentence that
Conversely, drift and glide only express motion through space, in the
is virtually equivaIent to the preceding two-clause sentence:
way that fIoatz does. They cannot aIso be used in a non-motion sense:
(9) The craft floated, into the hangar on a cushion of air *The canoe glided on that spot of the lake for an hour,

66 LEONARD TALMY Lexicalization pattems 6 7

Comparably, throw is semantically much like kick, in referring to a b. GET: [A,] INDUCE [AZ] t0 GO
distinct motion event caused by a prior body action: I rhrew rhe hall 1 GOT him out of his hiding place, by luringlscaring him
across the fieid wifh my fefi hand. But it has no usage parallel to kickl j1 lured/scared him out of bis hiding place.
referring to the body action abone - i.e., to swinging an object around Similarly: 1 talked him down off the ledge.
with one’s arm without releasing it into a separate path. By contrast
1 prodded the cattle into the pen.
swing itseIf is generally restricted to this latter sense, parallel to kickI,
C. URGE: [A,] AIM t0 INDUCE [AZ] t0 GO
and cannot be used in a sentence like *I swurtg rhe hall across rhefìeid
1 URGED him away from the building, by waving at him.
wirh my leji arm to express consequent motion through space.
31 waved him away from the building.
All these forms fit - and can further illustrate - the lexicahzation
formulas of (2) and (3). When plugged into (2), the forms immediately Similarly: 1 beckoned him toward me.
above exlribit not only usage equivalence but also semantic equivalerme. B. conflation involving Motion in other recurrent semantic
Thus, the usage and meaning of rhrow (LJ is the same as that of swing complexes
(LJ when this form is in construction with the largely grammatical d. GIVE: [A,] AMOVE [F] into the GRASP of [AZ]
sequence (G) causero move by . . . -ing (‘throw’ = ‘cause to move by
1 GAVE him another beer, sliding it
swhging’). And as for kick, thls form is seen to possess a range of usages
$1 slid him another beer.
because it can be plugged into borh sides of formula (2): kick2 = cause ro
move by kicking,; or, equivalently by formula (3), kick (Ll) has usages e. PLACE: [A] A~~~~ [F ro G] with limb motion but without
equaling the usage of rhrow (Ll) taken together with the usage of swing body translocation
04).‘” 1 PLACED the painting down on the table, it lying there.
In the languages that have it, the conflation pattem being descrlbed $1 laid the painting down on the table.
here normally applies far beyond the expression of simple Motion. It Similarly: 1 stood/leaned/hung the painting on the
extends as well to Motion compounded with mental-event notions chairlagainst the door/on the wali.
(IRA), to Motion compounded with other specific material in recurrent f . COVER: [F] BE‘ all-over [G]
semantic complexes (I~B), to embeddings involving more than one
Paint COVERED the rug, BEING in streaksldots
Motion event (r3c), and to metaphoric extensions of Motion (13~).
+ Paint streaked/dotted the rug.
Below, small caps indicate a posited ‘deep’ or ‘midlevel’ morpheme, one
that represents a basic semantic element or a specified semantic C. double conflation: example of a lexical triplet
complex. As an underlying main clause verb, it conflates with an g . Could you GTVE me the flour,
element (usually the verb) from the accompanying subordinate clause. having first AMOVED it down off that shelf,
Again, virtually none of these additional forms can be expressed as such
having first reached, to it with your free hand?
in languages like Spanish.
* Could you GIVE me the flour,
(13) Extensions of the Motion conflation pattern in English having first reached, it down off that shelf with your
(F = Figure, G = Ground, A = Agent, (to) AGENT = (to) cause free hand?
agenthely, AMon= agentively Cause t0 Mo%, Sr&1 Caps=a
j Could you reach, me the flour down off that shelf with
deep or midlevel morpheme)
your free hand?
A. conflation involving Motion compounded with mental- D. confiation involving metaphoric extensions of Motion
event notions non-agentive
a . GO: [A] AGENT himself [Le., his whole body, = F] to MOVE h. ‘MOVE': [F] MOVE metaphorically (i.e., change state)
She WENT to the party, wearing a green dress. He ‘MOVED' to death, from choking-on a bone.
3 She wore a green dress to the party. ($He died from choking on a bone. -or:)
Similarly: 1 read comics al1 the way to New York. +He choked to death on a bone.

68 LEONARDTALMY Lexicalization patterns 6g

i. BECOME: in the environment: Adjective
'MOVE' or Cause is expressed in the same sentence, it must be as an indepen-
dent, usually adverbial or gerundive type constituent. In many Ian-
The shirt BECAME dry, from flapping in the wind.
guages - for example Spanish - such a constituent can be stylistically
(+The shirt dried from Aapping in the wind. --OK)
awkward, so that information about Manner or Cause is often either
3 The shirt flapped dry in the wind.
established in the surrounding discourse or omitted altogether. In any
Similarly: The tinman rusted stifE. case, it is not indicated by the verb root itself. Rather, languages of this
The coat has wom thin in spots. type have a whole series of surface verbs that express motion along
The twig froze stuck to the window. various paths . l l This conflation pattem can be represented schematicalIy
j. FORM: [F] 'MOVE' into WSTENCE (cf. the phrase come into as in Figure 2.2.
existence)
A hole FORMED in the table, from a cigarette buming it. Figure Motion Path Ground
3A hole bumed in the table from a cigarette.
agentive
k. 'AMOVE': [A] AGENT [F] t0 'MOVE'
1 ‘*MOVED' him to death, by choking him.
<surface verbss
($1 kilIed him by choking him. -or:)
+ I choked him to death. Figure 2.2 Path conflated in the Motion verb
Similarly: 1 rockedlsang the baby to sleep. Language families that seem to be of this type are Semitic, PoIyne-
1. *BECOME = MAKE~: ‘AMOVE' in the environment:-Adjective sian, and Romance. Spanish is a perfect example of the type. We draw
1 MADE~ the fence blue, by painting it. on it for illustration, first with non-agentive sentences, and point out
+ I painted the fence blue. how pervasive the system is here:
m. ~Fom = MAKE~: [A] AGW [F] to ‘MO& into EXISTENCE (~3.
the phrase bting inro existence) II51 Spanish expressions of Motion (non-agentive) with conflation
of Path
1 MADE~ a cake out of fresh ingredients, by baking them.
+I baked a cake out of fresh ingredients. a. La botella entró a la cueva (flotando)
Similarly: I knitted a sweater out of spun wool. the bottle moved-in to the cave (floating)
I hacked a path through the jungIe. ‘The bottle floated into the cave’

Mandarin Chinese is the same type of language as English. It b. La botella salió de la cueva (flotando)
codates Manner or Cause with Motion in its verbs. But the parallel the bottle moved-out from the cave (floatig)
goes further. It also has the same double usage for a single verb form: ‘The bottle floated out of the cave’

(14) a. Wõ yòng zu6 ji50 ti1 le yi xià qí=% c. La botella pasú por Ia piedra (flotando)
1 use(-ing) Ieft foot kick PERF one stroke waI1 the bottle moved-by past the rock (floating)
‘I kicked the Wall with my Ieft foot’ ‘The bottle floated past the rock’
b. W6 yòng zu6 jião bã qiú ti2 g u ò le cão-ch8ng d. La botella pació Po’ el tubo (flotando)
1 use(-ing) left foot OBJ ball kick across PERF field the bottle moved-through through the pipe (fIoating)
‘1 kicked the ball across the field with my left foot’ ‘The bottle floated through the pipe’

1.2 Motion+P& e. El globo subió Por la chimenea (flotando)
In the second typological pattem for the expression of Motion, &e verb the balloon moved-up through the chimney (floating)
root at once expresses both the fact of Motion and the Path. If Manner ‘The balloon floated up the chimney’

70 LEONARD TALMY Lexicalization patterns 71

f. El globo bajó Por la chimenea (flotando) ( or: Retorcí e l corcho y lo saqué -de l a b o t e l l a )
the balíoon moved-down through the chimney (Aoating)
1 hvisted the cork and it I *moved out from the bottle
‘The balloon floated down the chimney’
‘1 twisted the cork out of the bottle’
g. La botella se fié de la orilla (flotando)
the bottle moved-away from the bank (floating) And we can see it for Cause:
‘The bottle floated away from the bank’ c. Tumbé el árbol serruchandolo / / a hachazos / con
h. La botella volvió a la orilla (flotandoj 1 felled the tree sawing it by ax chops with
the bottle moved-back to the bank (floating) una hacha
‘The bottle floated back to the bank an ax
i. La botella le dió vuelta a la isla (flotando) ‘1 sawed // chopped the tree down’
the bottle to-it gave tum to the island (floating) d. Q&é el papel del paquete cortandolo
(= ‘moved-aroundl) I *moved off the paper from the package cutting it
‘The bottle floated around the island ‘1 cut the wrapper off the package’
j. La botella cruzó el canal (flotando) Among such agentive forms, those that refer to the placement or
the bottle moved-across the canal (floating) removal of a Figure object - the ‘putting’ verbs - together comprise a
‘The bottle floated across the canal subsystem that, again, invoIves different verb forms for the separate
k. La botella iba por el canal (flotando) indication of distinctions of Path, as seen in Table 2.1.
the bottle moved-along along the canal (floating)
‘The bottle floated along the canal Table 2.1. Spanish ‘putting’ verbs, differing according to distinctions of
Path (A = Agent, F = Figure object, G = Ground object)
1. La botella andaba por el canal (flotando)
the bottle moved-about about the canal (floating) A poner F en G A PU? F OIltO G
‘The bottle floated around the canal A meter F a 12 A PUt FiIltO G
A subir F a G A PUt F U.p (0ll)tO G
m. Las dos botellas se juntaron (flotando) A jUIltX F1 & F2 A PUt F1 & F2 tOgethE-1
the two bottles moved-together (floating) A qUitar F de G AtakeFOffG
‘The two bottles ffoated together’ A sacar F de G A take F OUt Of G
A bajar F de G ~take F down from G
n. Las dos botellas se separaron (flotando) A separar F1 & Fo ~take F1 & F2apart
the two bottles moved-apart (floating)
‘The two bottles floated apart’
Notice that English does use different verb forms here, pur and take,
In its agentive forms as well, Spanish shows the same pattem of in a way that at first suggests the Spanish type of Path incorporation. But
conflating Path in the verb. Again, Manner or Cause, if present, is an altemative view is that these are simply suppletive forms of a single
expressed in an independent constituent. We can see this for Manner: more general and non-directionai ‘putting’ notion, where the specific
form that is to appear at the surface is determined completely by the
(16) Spanish expressions of Motion (agentive) with conflation of
particular Path particle and/or preposition present. This single ‘putting’
Path notion involves an Agent moving a Figure with respect to a particular
a. Merí el barril a la bodega rodandolo location by the action of some body par-t(s) without the whole body
1 Amoved in the keg to the storeroom rolling it moving through space (and, hence, is distinct from the ‘carrying’ notion
‘1 rolled the keg into the storeroom’ of cany/take/bring). In expressing this notion, English uses pti in
b. Saqué el corcho de la botella retorciendolo conjunction with a Yo’-type preposition (1 put the disk into/onto the
1 Amoved out the cork from the bottle twisting it stove), take with a ‘from’-type preposition excépt when up is present (1

LEONARD TALMY Lexicaiization p a t t e r n s 73
72

took the dish off,fout of the stove), pick with a ‘from’-type preposition in Figure Motion Path Ground Manner
the presente of up (1 picked the dish up off the chair), and move with an I Cause 1
‘along’-type preposition (1 moved the dish further down [he ledge). As move
further evidente of their purely formal character, these distinctions of 7’ 1, 1.be,
verb form are effaced when there is Manner conflation. Thus, beside a
diierent-verb pair of sentences such as Iput the cork into/took the cork <sudace verbs>
out of the bottle is the same-verb pair I twtited the cork into/out of the Figure 2.3 The Figure conflated in the Motion verb
bottle, where the Manner verb rwist supplants both put and take.
Comparably, beside I put the hay up onto/took the hay down off the
platform is I forked the hay up onto/down off the platform. Thus, it can a. It rained in through the bedroom window [non-agentive]
(17)
be seen that any Path information borne by the English ‘putting’ verbs is
b. 1 spat into the cuspidor [agentive]
less than and no different from that expressed by the partides and
prepositions occurring in the same sentence and, accordingly, they can
be readily supplanted under the Manner conflation typical of English. But in the languages for which this pattem is characteristic, there are
On the other hand, the Spanish ‘putting’ verbs express the bulk of Path scores of Motion + Figure verbs with the most colloqnial and extensive
distinctions - the only prepositions used with this subsystem are a, de, of usages. Atsugewi, a Hokan language of northem California, is an
and en - and so are central, unsupplanted fuctures in the Spanish example par excellence of this type. The following verb roots are just a
sentence, as is typical for thczt Iangnage. sampling:
English does have a certain number of verbs that genuinely incor-
porate Path, as in the Spanish cotiationtype, for example: enter, exit,
(18) Atsugewi verb roots of Motion with conflated Figure
pass, rise, descend, return, circle, cross, separare, join- And these verbs
-hp- ‘for a small shiny spherical object (e.g. a round candy,
even calI for a Spanish-type pattem for the rest of the sentence. Thus, an eyeball, a hailstone) to move/be-located
Manner must be expressed in a separate constitnent as in The rock *;* ‘for a smallish planar object that can be functionally
passed by our tent cin its sZide/in siiding) by contrast with the usual
affíxed (e.g. a stamp, a clothing patch, a button, a
English pattem in The rock slid past our rent. But these verbs (and the
shingle, a cradle’s sunshade) to movejbe-located
sentence pattem they call for) are not the most characteristic of English. -caq- ‘for a slimy lumpish object (e.g. a toad, a cow drop-
In fact, the majority (here al1 except rise) are not original English forms ping) to move/be-located
but rather borrowings from Romance, where they are the native type. -swal- ‘for a limp linear object suspended by one end (e.g. a
shirt on a clothedine, a hanging dead rabbit, a flaccid
penis) to move/be-located
-qput- ‘for loose dry dirt to move/be-located’
1.3 Motim + Figure -sta+ ‘for runny icky material (e.g. mud, manure, rotten
In the third major typological pattem for the expression of Motion, the
tomatoes, guts, chewed gum) to move/be-located
verb expresses the fact of Motion together with the Figure. Languages
with this as their characteristic pattem have a whole series of surface
verbs that express various kinds of objects or materials as moving or qese verb roots can also have an agentive meaning. For example,
located. This conflation type can be represented schematically as in -staq- has the further meaning option: ‘(for an Agent) to move runny
Figure 2.3. icky material’. Thus, such verb roots typically function equally in the
This pattem can !Zrst be ilIustrated close to home, for English does expression of events of location, of non-agentive motion, and of
have a few forms that conform to it. Thus, the non-agentive verb (to) agentive motion. Each of these usages is now exemplified with -si&-,
rain refers to rain moving, and the agentive verb (to) spit refers to here in referring to guts (an instance of ‘runny icky material’). Each
causing spit to move, as seen in (17). example gives both the morphophonemic and the phonetic form (the

74 LEONARD TALMY Lexicalization parterns 7 5

superscript vowel represents a special morphophoneme of this lan- Table 2.2. The three typological categories for Motion verbs
gua&: Language/Iangtuge family The particular cumponents of a Motion event
(19) Atsugewi expressions of Motion with conflated Figure characteristically represented in t h e verb
root
a. locative sufiix: -Ík* ‘on the ground’
instrumental prefix: uh- ‘from “gravity” (an object’s own Romance Path + fact-of-Motion
SemitiC
weight) acting on it’ PoIynesian
a
inflecíional affix-set: ‘- w- - ‘Sd person subject (factual Ne2 Perce
mood)’ Caddo
/‘-w-uh-Sta+--“/ * [+o&a$k-a] Indo-European (all?) except Romance MannerjCause + fact-of-Motion
Literal: ‘Runny icky material is located on the grocnd from Chinese
its own weight acting on it’ Atsugewi (and apparently most Figure + fact-of-Motion
Instantiated: ‘Guts are lying on the ground northern Hokan)
Navajo
b. directional su&: -i& ‘into liquid
instrumental preflx: ca- ‘from the wind blowing on the
Figure’
a with the Motion verb to form any language’s core system for expressing
inflectional aft?x-set: ‘- w- - ‘Sd person subject (factual
mood)’ Motion. Conflations of this sort may not even form any minor systems.
/‘-w-ca-sta+&/ * [iwastaií&a] Sporadic instances of such a conflation do occur, however, and can
provide an idea of what a larger system might be like. The verb root
Literal: ‘Runny icky material moved into liquid from the
-piane iu the (Ameritan) English verbs emplane and deplane can be
wind blowing on it’
Instantiated: ‘The guts blew into the creek taken to mean ‘move with respect to an airplane’, that is, to specify a
particular Ground object plus the fact of Motion, without any indication
c. directional suffix: 4s ‘into fire’ of Path. It is the separate prefixal morphemes here that ‘specify
instrumental prefix: cu- ‘from a linear object, moving axial- particular Paths. What a ful1 system of this sort would have to con& is
ly , acting on the Figure’ the provision for expressing many further Paths, say, as in circumplune,
infIectionai af&set:s- ‘- w- -B ‘1 - subject, Sd person object ‘move around an airpIane’, and rransplane, ‘move through an airplane’,
(factual mood)’ as well as many further verb roots that participated in such formations,
/s-‘-w-cu-sta&is-a/ j[s&siáqcha] say, (fo) heme ‘move with respect to a house’, and (ro) liquid, ‘move
Literal: ‘I caused.it that runny icky material move into fire with respect to liquid’. But such systems are not to be found. It is not
by acting on it with a linear object moving axially’ clear why the Ground component should be so disfavored. Gne might
Instantiated: ‘1 prodded the guts into the fire with a stíck’ first speculate that, in discourse, the Ground object of a situation is the
most unvarying component and therefore the one least needing speci-
1.4 Manner/Cause, Path, ami Figure in a typology for Motion verbs fication. But on further consideration, the Figure would seem to be
The three basic conflation patterns for Motion verbs that languages even more constant, yet it forms the basis for a major typological
exhibit, in an apparently exhaustive typology, is summarized in Table system. One might next speculate that the Ground object is the
2.2. Subcategorization of these three types, based on where the component least salient or accessible to identification. But there seems
remaining components of a Motion event are expressed in a sentence, is nothing more obscure about airplanes, houses and liquids (to pick some
treated later. likely Ground objects) than, say, about notions of Path, which do form
Because it is apparently exhaustive, this typology raises questions the basis for a major typological system.
about the non-occurring combinatory possibilities. It can be seen that Explanation may next be sought in a concept of hierarchy: the
one Motion-event component, the Ground, does not by itself conflate different conflation types seem to be ranked in their prevalence among

76 LEONARD TALMY Lexicalization pattems 77

the world’s languages, with conflation of Path as the most extensively x.5 Aspect
represented, of Manner/Cause next, and of Figure least so. It may ‘Aspect’ can be characterized as the ‘pattem of distribution of action
therefore be the case that Ground conflation is also a possibility, but one through time’. The term ‘action’ as used here applies to a static
so unlikely that it has not yet been instantiated in any language (that has condition - the continuance of a location or state - as well as to motion
come to attention). However, while great disparity of prevalence for the or change. In Figure 2.4 are some of the aspe&types lexicalized in verb
different conllation types would be most significant if proved by further roots, with both non-agentive and agentive English verbs exemplifying
investigation, it would then itself require explanation, so that the each.
present mystery would only have moved down a Ievel. a. b. d. e.
one-way one-way Ll- muItiplex steady- Lient
There are further combinatorial possibilities to be considered. non-resettable resettable cycle state
Among these: rwo components of a Motion event conflating with
fact-of-Motion in the verb root. Minar systems of such conflation do
exist. For example, the Ground and Path together are conflated with
Motion in a minor system of agentive verbs in English, with forms like
shelve ‘cause-to-move onto a shelf (1 shelved the bookr) and box
‘cause-to-move into a box’ (1 boxed the apples). (The particular Paths die fall flash breathe d=P widen (intrans)
kti dmp bit beat -J-V widen (trans)
occurring in this system appear to be virtually limited to the contact-
forming ‘into/onto’ type; exceptional, thus, is quarry ‘cause-to- Figure 2.4 AspectuaI meaoings lexicalized in verb roots
move out of a quarry’, as in We quurried the granite, and the verb
mine with a similar sense, We mined rhe bauxite.) Another minor Various grammatical tests demonstrate the distinctness of these types
system of agentive verbs in English conflates the Figure and Path and of the verb roots incorporating them. The resettable type of a
together with Motion: powder ‘cause facial powder to move onto’ one-way verb is distinguished from the non-resettable type by its
i (She powdered her nose), scale ‘cause the scales to move off of (1 scaied compatibility witb iterative expressions, as in He fell 3 times; the
the fih). non-resettable verbs carmot occur here: *He died 3 times. TC same
Conflation systems of this multi-component sort apparently never one-way form is distinguished from a full-cycle form by its ability to
fonn a language’s major system for expressing Motion. The reason for appear in sentences like He fell and rhen got up, which the latter cannot
such a prohibition seems straightforward for systems observing finer do: * The beacon flashed and rhen went off. A gradient verb can appear
semantic distinctions: these would entail an enorrnous lexicon. There with adverbs of augmentation, as in The river progressively widened,
would have to be a distinct lexical verb for each fine-grained semantic unlike a steady-state verb: *She progressively siept. And so on,
combination - for example, beside box meaning ‘put into a box’, there Sometimes al1 that distinguishes two verb forms which otherwise have
would have to be; say, a verb foo ‘take out of a box’, a verb baz ‘move the same core meaning is a difference in incorporated aspe& In cer-
around a box’, etc., and further verbs for the myriad of Ground objects tain sectors of their usage, tbis is the case with Zeam, which (for many
other than a box. Such a system would be infeasible for language, whose speakers, though not for all) incorporates a completive aspect, and
organization relies less on large numbers of distinct elements and more study, which is steady-state. The semantically comparable verb teach has
on combinatorial devices that operate with a smaller set of elements. a lexicalization range covering both of these aspect-types:
However , one can imagine another kind of multi-component conflation-
(20) completive aspect steady-state aspect
al system, one with fairly broad-band referentes and hence fewer total We leamed/*studied French We *leamed/studied French
elements, aciing as a kind of classificatory system, that contained verbs in 3 years for 2 years
with meanings like ‘move to a rounb object’, ‘move from a round She taught us French in 3 She taught us French for 2
object’, ‘move throughlpast a round object’, ‘move to a linear objeci’, years years
‘move from a linear object’, etc. A system such as this would indeed be
feasible for language, yet also seems prohibited, and an explanation Lexicalized aspect figures in the analysis of a language in several ways.
here, too, must be awaited. First, aspect generally seems to be part of the intrinsic meaning of verb

78 LEONARD TALMY Lexicalization patterns 7 9

roots.r2 It is doubtful that any verb root can have a meaning wholly (21) Di8erent types of causative meaning incorporated in the verb
neutral to aspect - even in Ianguages where the root is aiways sur- root
rounded by aspect-specifying inflections. a. The vase broke - autonomous event
Second, a verb root’s intrinsic aspect determines how it interacts with (not causative)
grammatical elements that also have aspectual meaning. Many of the
b. The vase broke from a - resulting-event causation
Iatter appear only with verb roots of a particular aspect-type, operating
ball’s rolling into it
on them to yield a different aspe&type as a resultant. For example, in
English the grammatical forrn keep -ing operates on a one-cycle verb of c. A balI’s rolhng into it - causing-event causation
the (c) type to yield a mu1tiplex aspectual meaning of the (d) type. This broke the vase
shift takes place for flash in The beucon keprflashing. Simi1arly, we can d. A hall broke the vase - instrument causation
make the reverse change from the (d) type to the (c) type with the (in roliing into it)
abstract grammatical form VdWY c1 [-+ Deriv], - that is, by using a e. I broke the vase in rolling - author causation (i.e. with
construction that has the verb root in a derived nominal form. This is a hall into it result unintended)
what happens to the verb root brearhe (with an inherent multiplex
f. 1 broke the vase by rolling - agent causation (i.e. with
meaning) in the sentence She took a breath (with a ‘once only’
meaning) . l3 a ball into it result intended)
Third, different languages have different patterns of aspect incorpora- g. 1 broke my arm when I fell - undergoer situation (not
tion in their verbs. For example, we will see in Section 1.7 how verbs (= My arm broke [on me] causative)
referring to states are Iexicalized in some languages with the (b) when I fell)
‘ene-way’ aspect-type - with the sense of entering into the states - whiIe h. 1 walked to the store - self-agentive causation
for the same states other languages wiI1 use the (e) ‘steady-state’ i. 1 sent him to the store - inductive causation (caused
aspect-type. And fourth, aspect incorporation can correlate with sur-
%wCY)
rounding factors. For example, it seems generally that a language with a
ready ingection indicating ‘multiplexity’ has few verb roots like Engiish
bear, wag,flap, brearhe with inherent multiplex aspect. Rather, the verb The autonomous (a) type presents an event as occurring in and of
roots by themselves refer to one cyde’s Worth of the action, and take the itself, without impiying that there is a cause (such causes as there may be
inflection to signal multiplexity. One language apparently like this is fa11 outside of attention). lJ In the (b) ‘resulting-event causation’ type, on
Hopi (Whorf r956), and another is Ameritan Sign Language (Ehssa the other hand, this main event has resulted from another event
Newport, personal communication). (expressed in a subordinate ciause or a nominalization) and would not
otherwise have occurred. English verbs that incorporate both these
causation types but no others are die, fall, dri;fr, disappear, sleep.
1.6 Causarion While the (b) type focuses on the main event as res&@ from another
By one andysis, there are quite a few distinct types of causation event, the (c) Icausing-event’ type focuses on the 1atter (now the
lexicalized in verbs (see Talmy r976a). The number is appreciably subject) as causing the main event.i6 And the instrumental (d) type
greater than the usually recognized two-way distinction between ‘non- focuses on just that object within the causing event that actually
causative’ and ‘causative’. Some verbs incorporate only one causation impinges on the affected elements of the resulting event.” English has
type while others demonstrate a range of incorporations. A number of very few verbs that incorporate the (c) or (d) types without also
such types are listed below, in order of increasing complexity or incorporating the (e) and (f) types. One example, though, is erode as in
deviation from the basic (except for the interposed type of (zrg)). AI1 The river’s rushing along it/ The river/ ? * The sciemists eroded that section
but two of these types can be ilIustrated with the verb break; other verbs of land.
are given to illustrate rypes (h) and (i). Most of these types are here In both author (e) and agent (f) causation, an animate being wills a
named for the kind of element that acts as the verbal subject:14 bodiIy action that leads (through a variously sized chain of causal

80 LEONARD TALMY Lexica~ization patfems 81

events) to the main event referred to.18 In the author type, the being b. The boy rolled the log across the field
intends all these events except the final one; in the agent type, the final - agent causation
one, toa, is intended. English verbs associated with the author type and c. The boy roIled across the fieId on purpose
only slightly or not at al1 with the agentive are spili, drop, knock (down), - self-agentive causation
and bimorphemic mislay. Strictly agentive verbs are murder, fhrow,
persecute. In the inducive (i) type, something (whether a thing, an event, or
The undergoer in the (g) type is Iike an author in that he does not another Agent) induces an Agent to intentionally carry out an act.19
intend the event mentioned. But he also has not intentionally under- Some English verbs incorporating this type are: send, drive (off), chase
taken any actions that culminate in that event. Rather, the event is (fzway), smoke (out), lure, aftract, repel, sic . . . on. The verb set . . .
conceived of as occurring independently of the undergoer, but as upun has a range that permits a contrastive example:
affecting his subjective state, usually adversely. Many languages express (25) a. The dogs set upon us - se&agentive causation
the undergoer in an oblique constituent, as does Spanish:
b. He set the dogs upon us - inducive causation
(caused agency)
(22) Se me quebró el brazo Our method for distinguishing causation types rests on fmding verbs
‘The arm broke itself [to] me’ = ‘1 broke my arm’
that incorporate only one type or that have ranges differing by only one
Se me perdió la pluma
type (or, at least, ranges which overIap in enough different Ways). For
‘The pen lost itself [to] me’ = ‘I lost my pen’
example, we can try to use each of the verbs die, kiZl, murder in every
one of the causative types listed in (21):
English does have this construction (with on: My arm broke on me). But
it also has verbs that alIow the undergoer as subject (1 broke my arm, 1 (26) a. He died/*killed/*murdered yesterday (i.e.: ‘He underwent
caz@ my sweater on a nail, I developed a wart in my ear) as well as death’)
ones that require it that way, like lose andforger. We can contrast the 6. He died/* kiIled/*murdered from a car hitting him
agent , author, and undergoer types with the tbree verbs in 1 c. A car’s hitting him *died/killed/*murdered him
hid/misZaid/losr my pen somewhere in the kitchen. Tbese verbs al1 have
d. A car *died/killed/*murdered him (in hitting him)
a similar core meaning, one involving an object’s becoming not findable.
But each incorporates a different causation type: e. She unintentionally *died/killed/*murdered him
f. She *died/killed/murdered him in order to be rid of him
approx. g. He *died/*killed/*murdered his plants (ie.: ‘His plants
that NP become = died on him’)
(23) not-frndable
t0 UNDERGO h. He *died/*killed/*murdered (ie.: ‘He killed himself by
intemal will’)
The self-agentive (h) type is like the agentive except that the animate i. She *died/*killed/*murdered him (i.e.: ‘She induced him to
being’s bodily action is itself the final and relevant event, not just a kill [others]‘)
precursor. Often, the whoIe body is moved through space as a Figure. In From (26) we can derive the summary in TabIe 2.3 where we see just
their usual usage, the English verbs go, walk, run, jzunp, trudge, recline, the acceptable usages. From the different acceptability pattems here,
crouch, etc., incorporate this type. The verb rol2 can incorporate severa1 we can establish that the agentive (f) is a type by itself (it alone
different causation types, among them the self-agentive, and so permits accommodates murder) and that there are at least distinctions between
a contrastive example: the (a/b) set of types (die but not kill ranges over these), the (c/d/e) set
(24) a. The log rolled across the field of types (kiD’s range minus the agentive (f), already isolated), and the
- autonomous event (g/h/i) set of types (suiting none of the verbs). We can now seek cases

Lexicaiization pattems 8 3
82 LEONARD TALMY

b. I intentionally ‘midaid/hid my pen somewhere in the
TalAe 2.3. fypes of cu~ative mage: die, kill and murder
kitchen
die kill murder 1 *mislaid/hid the pen so that it would never be seen again
1 intend to *mislay/hide my pen somewhere in the kitchen
She persuaded me to *mislay/hide the pen
: ;
C J
*Mislay/Hide your pen somewhere in the kitchen!”
d dT
Table 2.4. Lexicaiized causation types shifted by grammatical elemena
f $ J
autonomous agentive self-agentive undergoer inducive
h v- make v
i ib’, {;-make wLv
or v) , have v
71 VBVREFL
e Iv 0r VI , have v

that exhibit distinctions within these clusters of types. The (g) type can (a)-(e) correspond to (a j(e) in (29).
be separated out by the fact that it alone accommodates the verb lose (in
its ‘not findable’ sense), as we could demonstrate with an array of Further evidente that verbs have different causative lexicalizations is
sentences like that above. Besides, (g) has already been distinguished that they take different grammatical augments to indicate a shift in
from (h) and (i) in that break can incorporate it but not the latter two causation type. TabIe 2.4 shows a sampIe from English of such augments
types. These themselves are distinguished in that only (h) accommo- and the shifts they mediate. In (zg) each shift is illustrated with a verb
dates trudge and only (i) accommodates sic , . . un. And so on. that is lexicalized solely in the starting-point causative type and is placed
We can establish more conclusively that a verb incorporates a with the relevant grammatical shifters in a clause. Accompanying this,
particular causation type by using special test frames. For example, here for comparison, is a causatively equivalent clause with an unaugmented
are two sets of frames that can test for author- and agent-type verb (in italics) lexicalized solely in the causation type at the end of the
incorporation in English verbs: shift. Thus, (2ga) shows disappear, which is solely autonomous (The
stone disappeared/* The witch disappeared the stone), rendered agentive
s: author-causative s: agent-causative by the augment make, and thereby equivalent to the unaugmented
(27) obliterare, which itself is solely agentive (*The stone obliteruted):U
s accidentally s intentionally
s in (+ Cause clause) s in order that . . . (29) a. The witch made the stone (cf. The witch obliterated the
S . . . toa . . . NP intend to s
disappear stone)
may s! NP~ persuade NQ to s
s! b. He made himself disappear (cf. He scrammed)
c. You might have your toy (cf. You might lose your
sailboat drift off toy sailboat)
When placed in these frames, the verbs mislay and hide show com-
You might have pour wallet (cf. You might lose your
plementary acceptability patterns. In this way each verb is shown to (get) stolen in the crowd
incorporate the one but not the other of the two causation types tested wallet in the crowd)
for:21 d. She dragged herself to work (cf. She hudged to work)
e. I had the maid go to the (cf. I senr the maid to the
a . 1 accidentally mislaid/*hid my pen somewhere in the kitchen store store)
(28)
1 mislaid/*hid the pen in putting it in some obscure place 1 had my dog attack (cf. 1 sicced my dog
May you mislay/*hide your pen! the stranger on the stranger)

84 LEONARD TALMY Lexicalizatioti p a t t e r n s 8 5

We can observe causative lexicalization patterns at different levels of agentive in Spanish. Japanese adds an inflection to its verbs to express
linguistic organization. At the leve1 of individual Iexical items, a verb’s the corresponding agentive, while Spanish adds its reflexive clitics (here
particular range of lexicalizations can often be explained on the basis of serving not in a ‘reflexive’ but in a ‘de-agentivizing’ function) to express
its core meaning alone. For example, the basic referent of break can the autonomous. We can illustrate these complementary pattems with
apply to a person’s body-part but not to his whole body (1 broke bis the verbs for ‘open’:
arrn/ *I bruke him) and, accordingly , the verb Iacks a self-agentive usage
(*I broke, in the sense ‘1 broke myself/my body’). Similarly, erode @) Japanese: a. Doa ga aita
resists agentive usage because an agent cannot generally marshal the door SUBJ open(PAsT)
instrumentalities of erosion. On the other hand, it seems purely ‘The door opened
arbitrary that poison has an agentive but not an autonomous usage (He b. Kare wa doa o aketa
pononed her with toadstools/*She poisoned after eating toa&toob) he TOP door 0~ open (taus PASAST)
while drown has both (He drowned her/She druwned), or that conceal ‘He opened the door’ I
has an agentive but not a self-agentive usage (1 concealed her/*She Spanish: c. Abrió la puerta
concealed in the bushes) while hide has both (1 hid her/She hid in the he opened the door
bushes). But motivated or idiosyncratic, all these !exicalization pattems ‘He opened the door’
are associated with particular lexical items.
d. L,a puerta se abrió
There are also pattems operating at the leve1 of a whole semantic
The door REFL opened
category. For example, virtudly al1 English verb that refer to death
without expressing its cause (in contrast, for example, to drown) ‘The door opened
observe the basic causative/non-causative distinction - i.e., are lexical- Finally, at the broadest scope, some lexicalization pattems affect the
ized for either the non-causative (zr a/b) types or the (21c-e) causative whole lexicon of a language. One example is that in Japanese the
types but not for both. The pattern applies to both simplex and complex causing-event (21~) and instrument (2Id) causation types are barely
expressions: represented at all. Thus, verbs otherwise corresponding to our kill and
(30) non-causative causative break cannot be used (without extreme awkwardness) with the causing
event or Instrument as subject. To express these constituents, one must
die kick off kill exterminate use the (2Ib) resulting-event causation type instead.
expire kick the bucket slay Off
decease bite the dust dispatch waste
perish give up the ghost murder knock/bump off I .7 Interactin of aspect and caus&n
croak meet one’s end liquidate rub out Different verb roots incorporate different combinations of aspectual and
pass away breathe one’s last assassinate do in causative types. One might at first expect a lauguage to have a roughly
slaughter do away with equal distríbution of the combinations over its lexicon and to have
grammatical elements for getting from each combination to any other.
By contrast, almost all English verbs expressing the material disruption But we find two limiting factors. First, not al1 aspect-causative combina-
of an object - e.g. break, crack, snup, burst, bust, smash, shatter, shred, tions are relevant to every semantic domain. For example, in many
tip, lear - apply equally in both non-causative and causative cases (The Ianguages the semantic domain of Mates’ seems to involve only (or
balloon burst/l burst the balluon). There are not many more exceptions mainly) these three aspect-causative types (compare Chafe 1970):
than collapse, lacking an agentive usage (*I collapsed the shed), and
demoiish, lacking the autonomous usage (* The shed demoiished). a. being in a state (stative)
(34
Different languages often exhibit different lexicalization patterns for
b. entering into a state (inchoative)
a particular semantic category. For example, verbs referring 10 states are
mostly lexicalized in the autonomous type in Japanese but are mostly c. putting into a state (agentive)


Second, even for such a smaller set, the relevant verbs in a language b. Boku wa tatte ita
generally are not evenly lexicalized over the different types. For 1 TOP having-arisen was
example, for the expression of ‘states’, there are languages in which the ‘1 was standing’
verb roots are preponderantly lexicalized in only the (a) or only the (b) c. Hon o tateta
or only the (c) type. In other languages, such verb roots show a small book Om AGENTed-tO-atiSe
range of lexicalizations, either over the (a/b) types or over the (b/c) '1 stood the book up’
types. There are also Ianguages in which the same verb root is used
d. Kodomo o tataseta
equivalently for al1 three aspect-causative types. Sometimes a lan-
Child OBJ INDUCEd-to-tiSe
guage’s roots exhibit different pattems for different categories within
the ‘states’ domain. Wherever the verb roots are restricted in their ‘1 stood the Child up’
aspect-causative ranges, there are generally grammatical devices for Exemplifying the third pattem, Spanish lexicalizes posture notions in
getting to the remaining types. But because of all these limitations, the the agentive ‘putting-into-a-state’ type, the other types being derived
number of devices required can be quite small. therefrom. For example, the verb acosrar is inherently transitive, with
We first demonstrate these Iexicalization pattems for one category of the meaning Yo lay (someone) down’. To it must be added the reflexive
states, that of ‘postures’: postures or orientations that are assumed by morpheme, giving acostarse, to get the meaning Yo he down’. n And for
the human body or by objecis treated as comparable to the body.” We the steady-state meaning ‘to lie’, the verb must be stied with the past
can use English here to illustrate the pattern of lexicalization largeiy participle ending and put in construction with the verb ‘to be’: estar
limited to the ‘being-in-a-state’ type. This is seen in verbs like lie, sif, acostado?
stand, lean, kneel, squat, cruuch, bend, bow, etc.= These verbs must
(35) a. Acosté el niño b. Me acosté
generally take on additional elements for the other aspect-causative 1 laid down the Child myself I Iaid down
types to be conveyed. For example, lie by itself refers to being in the ‘I laid the Child down’ ‘1 lay down’
lying posture. The verb must be augmented by what we cal1 a ‘satelhte -
yielding the form lie down - to signify getting into the posture. And it c. Estaba acostado
must be further augmented by an agentive derivation - yielding lay 1 was laid down
duwn - to refer to putting into the lying posture? ‘I lay (there)’
(33) a. She lay there al1 during the program These typological findings can be represented together in a single
b. She lay duwn there when the program began schematic matrix, as in Table 2.5. For each class of language, Table 2.5
c. He laid her down there when the program began shows the aspect-causative type of the verb in which postura1 notions
are generally lexicalized, and the pattems by which the other types are
Japanese is a language where posture verbs are generally lexicalized derived therefrom.
in the ‘getting into a state’ type, with the other types derived therefrom.
For example, the basic meaning of tatu is ‘to stand up’ (comparable to Table 2.5. Lexicalization patterns for verbs of posture (v = ver& root,
the English verb arise). When this verb is grammatically augmentcd by SAT = satellite, PP = past participle inflection)
the -fe iru form, whose meaning can be rendered as ‘to be (in the state
be in a posture gel into a posture put imo a postwe
00 having [Ved]‘, the resultant meaning is ‘to be in a standing posture’.
And when the verb is augmented by the agentive or by the inducive English: V-V+SAT- VfCAUSfSAT

sti, yielding the forms tateru and tataseru, the resultant meanings are Japanese: %e'+v + PP -v + v+cAus

‘to put into a standmg posture’ a thing or a person, respectively. To Spanish: ‘be’ + v + PP’ VfREFL4 V
illustrate:
(34) a. Baku wa tatta Other Ianguages have other means for deriving the non-basic aspect-
1 TOP arose causative types from the favored one. For example, German is like
‘1 stood up’ English in having the stative type as basic for posture notions, as with

88 LEONARD TALMY Lexicalization patterns Sg

verbs Iike liegen and sifzen. But it does not derive the inchoative inchoative type. For the stative type, however, the grammatical form ‘be
‘getting-into-a-state’ type directly from this. Rather, it first derives the + past-participle-inflection’ must be added, yielding be frozen:
agentive ‘putting-into-a-state’ type, with verbal forms Iike legen and a. The water wfzs frozen
setzen. And from this, in the manner of Spanish, it uses tbe reflexive to (38)
b. The water froze
get back to the inchoative, with forms like sich legen and sich sefzen.
c. 1 froze the water
Schematically:
/ -% Tbe remaining possible two-way pattern - where the verb root would
(361 German: V V + CAUS + REFL - v+cAus
be used for, both the stative and the agentive types, but not the
In the preceding Iexicalization pattems, the verb root incorporated inchoative - does not appear to have any realization. One reason for
only one aspect-causative type. There are further pattems in which the such a lack may be that these two types do not share a factor that is
common to them but absent from the inchoative.
same verb form serves equally for two types, while grammatical
augmentation is required for the third. In one pattern of this sort, the These two-way cases bring us to the pattem where the same verb root
is used, without any grammatical augmentation, for al1 three aspect-
‘being-in-a-state’ and the ‘getting-into-a-state’ types are represented by
the same Iexical form, but an augmented form is used for the ‘putting- causative types. Tbis pattern seems to be the one English posture verbs
into-a-state’ type. The verb root in a pattem like this may be thought to are moving toward in a process of change going on now,2g and we can
capture a factor common to the two types it represents, namely, the see the pattem fully for severa1 individual verbs. of other ‘state’
involvement of only a single participant (note that the unrepresented categories. One clear example is hide, a ‘position’ verb:‘O
‘putting-into-a-state’ type, requiring an agent, involves two partici- a. He hid in the attic for an hour
c.39 1
pants). By one analysis, modem literary Arabic exemplifies this pattem - being in a position
for posture notions (but see below for an altemative interpretation), as
in the following root referring to ‘sleeping’ or ‘lying’: b. He hid in the attic when the sheriff arrived
- getting into a position
(37) a. Nãm-a t-@ll Calã c. 1 hid him in the attic when the sheriff arrived
the-Child-NoMon - putting into a position
{“zz} he {onto} shsearbed We can point to one further lexicalization pattem. Here, the verb
‘The Child was lying on the bed’/‘The Child lay down onto root is always accompanied by morphemes with their own aspect-
the bed’ causative meanings, making it difficult to determine whether the verb
b. Anam-tu t-tifl-a calà s-s& root itself incorporates any aspect-causative type of its own. Perhaps it
laid-down-1 the-Child-acc on the-bed does not, and the conclusion to be drawn is that such a verb root refers
‘I laid the Child down onto the bed solely to a particular state, abstracted away from al1 notions of aspect
and causation, and that it requires augmentation for every aspect-
In another pattem, the same verb root is used to express both the causative indication. Such augmenting morphemes can exhibit some of
inchoative ‘entering-into-a-state’ and the agentive ‘putting-into-a-state’ the same patterns of incorporation as seen above. In some cases, there
types, while a different formulation is required for the stative ‘being-in- would be distinct morphemes for each of the aspect-causative types. In
a-state’ type. The common factor captured by the verb with two usages other cases, a single set of elements would serve for some pair of
in this pattern would seem to be ‘change of state’. In familiar languages, aspect-causative types, with another set for the third. This Iatter pattem
there are no apparent instances of this as the predominant pattern for can be exemplified by Atsugewi. Here, a verb root referring to posture
verbs expressing postures. But if we switch here to another category of is always surrounded by aspect-causation indicating ties. And among
states, that of %onditions’ (further treated below), the pattem can be these, generally, one set serves for botb the ‘getting-into-a-state’ and the
exemplified by English. Here, for instance, the verb fiee2.e lexicalizes ‘putting-into-a-state’ meanings, while a different set is required for
the condition of ‘frozenness’ together with either the agentive or the ‘being-in-a-state’. This is illustrated in (40).


(401 a. verb root -itU- ‘for a linear object to be in Table 2.6. Lexicdization pattem for Latin verbs of condition (v = verb
// move into/out of/while root, PP = pmt pariiciple inflection)
in a lying posture’
directional su3ix: -II& ‘down onto the ground’ be in a condition enter into a condìtion put into a condition
infíectional affix-set: s- ‘- w- -a ‘1 - subject, 3d person object Y
Independent: V , V t INCHOATIVE V + CAUS
(factual mood)’ Y .
/s-r-w-itU-miE-a/* [&ithmíE] Dependent: ‘be’ + v + PP V+Mw~OPAsslvE~ V

‘1 lay down onto the ground’/‘I Iaid it down onto the ground’ Examples:

-it”- Independent: patere patescere patefacere
b. verb root: as for (a) above ‘to be open’ ‘to open (itr.) ‘to open (tr.)’
locative suffix: -ak- ‘on the ground Dependent: fractus esse frangi frangere
inflectional &-set: s- ‘- w- -a ‘I - subject, 3d person object ‘to be broken’ 70 break (intr.)’ ‘to break (tr.)’
(factual mood)’
/s-‘-w-itU-ak.-a/ j[s&it,ák*a] Second, a pattem in a language that predominates within one
‘1 was Iying on the ground’ category of a semantic domain may or may not do so ucro’oss the
Arabic forms like those cited earlier have an aitemative analysis that categories. As already seen, English is inconsistent here because its
places them at this point of the exposition. The verb root can be taken to posture verbs are generally lexicalized in the stative, while its condition
be a consonantal form that - like the Atsugewi root - names the state verbs have the two aspect-causative meanings other than stative.
alone and always takes different interposed vowel sequences as gram- Latin also exhibits dierent pattems across categories. To show this,
matical augmentations. These grammatical elements, then, follow a we first point out that what has so far been considered the single
pattem complementary to that of Atsugewi: one vowel sequence category of ‘conditions’ is better understood as comprising two separate
bandles both the stative and the inchoative, while another one handles categories. One of these is ‘independent conditions’: conditions that
the agentive . objects are conceived of as occurring in naturally. The other category is
that of ‘dependent conditions’: conditions conceived of as not original
I-7.1 Consistency of patiems within a lmguage for objects, enes that objects must be brought into by externa1 forces. In
Lexicalization pattems for aspect-causative types exhibit different de- many languages, independent conditions are frequently Iexicalized in
grees of pervasiveness in a language, first in the degree to which a adjectives. In Latin they are, too, but they also frequently appear in
pattem predominates within a semantic category. For example, posture verbs. Here they are generally lexicalized in the ‘being-in-a-state’ type,
notions in English are largely consistent in ti& stative lexicalization, with the other types derived therefrom. Dependent conditions, on the
with perhaps oniy inchoative atie falling outside this pattem. By other hand, are generally lexicalized in verbs in the agentive, and these
contrast, posture notions in Latin show up in verbs of a variety of follow the Spanish pattern for derivation (except that instead of the
lexicalization types. Each type of verb employs different means to yield reflexive, the mediopassive inflections are used), A schematic repre-
other aspect-causative meanings (e.g. stative sedere Yo sit’ takes a sentation is given in Table 2.6.
prefixal satellite to yield the inchoative considere ‘to sit down’, while The other languages we have looked at in this section show greater
agentive inclinare ‘to lean (something) against’ takes the reflexive to consistency across categories. They have the same Iexicalization pat-
yield the inchoative se inclinare ‘to lean (oneself) against’): tems for their verbs of condition as they do for their verbs of posare.
We illustrate this extension of the pattems first for Japanese (ea) and
(41) stative inchoative agentive Spanish (42b). Compare (34) and (35) with the following:
stare ‘stand’ surgere ‘stand up’ ponere ‘lay, set’
sedere ‘sit’ locare ‘set, lay’ ‘P:‘,anese
Y I Ya
iacere ‘lie’ inflectere .‘bow, bend h&u ga kootte ita Mim ga kootta Miiu 0 ko0ra.d
water suw frozen be (PA@ water SUBJ freeze (PAST) water OBT freeze (CAUSE PA@
cubare ‘lie’ inclinare ‘lean’ ‘The water was frozen’ ‘The water frote’ ‘1 froze the water’

LEONARD TALMY Lexicalization patterns 93
92
b. Spanish Y or prefklly:31 .
El agua e s t a b a h e l a d a El agua se heIb& el agua a. *He unhid from the attic
the water was frozen the water REFL froze 1 froze the water (46)
‘Tl-e water was froten’ ‘The water froze’ ‘1 froze the water’ b. *1 unhid him from the attic
Comparably, adjectives of condition have ready adjunct verbs or
Comparably, Arabic verbs referring to conditions are lexicalized like verb-foting affíxes to express state-location and state-entry but, in
posture verbs, with the stative and the inchoative using the same form. English and many other languages, not state-departure:32
Compare (37) with the foll&ving: be-in-a-state:
(47)
A’may-tu t-@-a be sick
(43) ‘Amiy-a
enter-into-a-state: exit-from-a-state:
was blind -he the-boy-NOM made blind-1 the-boy-Acc get sick *lose sick
became blind
‘1 blinded the boy’ sicken *desick
?he boy was/t>ecame blind’
put-into-a-state: remove-from-a-state:
makre (someone) sick *break (someone) sick
1.7.2 Other aspect-causative types sicken (someone) * desick (someone)
There are aspectxausative types other than the three listed in (32) that Ameritan Sign Language is similarly constrained. Thus, its signs for
might seem quite relevant to notions of states. These would invoive the conditions (like ‘sick’) can generally be executed with a number of
transition from being in a state to not being in that state. Such a distinct movement pattems indicating different aspects (‘be sick’, ‘be
transition could apply to both the non-agentive and the agentive: sick for a long time’, ‘stay si&‘, ‘becume si&‘, ‘become thoroughly
si&‘, ‘repeatedly become si&‘, ‘be prone to becoming si&‘, etc.), but
(~1 b’- exiting from a state c’. removing from a state state-departure is not among these (*‘cease being si&). The idea must
be expressed with a combination of two signs (‘be si& + ‘finish’).
However, such types of ‘state-departure’ seem to be under a universal It is not clear why there should be this avoidance of expressing
constraint excluding them from at least one type of lexicalization: a verb state-departure. But in any case, among grammatical elements it is only
root can refer to both state-location and state-entry, but it cannot refer a tendency, not an absolute. In Atsugewi, verb roots referring to
to one of these and also to state-departure. Thus, the Arabic verb form postures and positions (and apparently also conditions) regularly take
for ‘be/become blind’ cannot also mean ‘cease being blind’, and the grammatical elements that indicate state-departure, at least in the
English He hid can refer to ‘being in hiding’ or ‘going into hiding’, but agentive. We exemplify this with the verb root used previously in (40):
not also to ‘coming out of hiding’. Beyond such lexicaliiation ranges, -itu-
(448) verb root: ‘for a linear object to be in / /
‘tie exclusion of state-departure from lexicalization is total if it is move into/out of/ while in a
assumed that even singly lexicalized change-of-state verb roots (e-g. die) lying posture’
always refer to entry into a new state (‘death’) rather than departure
directional &Tix: -ii: ‘up off something’
from an old state (‘life’). inflectional affrx-set: s- ‘- w- -= ‘I- subject, 3d person object’
In addition, state-departure - though not excluded from them - seems /s-‘-w-it”-iE-a/ * [s&t-úi]
quite under-represented among grammatical devices that interact with
‘1 picked it up off the ground, where it had been lying’
verb roots. For example, English hide cannot be used with departure-
indicating satellites or prepositions, either in the postposed location: I .8 Persmalion
For actions of certain types, approximately the same actional content is
(45) a . *He hid out of the attic = He carne out of the attic, where he manifested whether one or two participants are involved. For example,
had been hiding whether John shaves himself or shaves me, the action still involves one
b. *I hid him out of the attic = 1 got him out of the a%c, where hand moving one razor over one face. The only relevant difference here
he had been hiding is whether the hand and the face belong to the same body. The

94 LEONARD TALMY Lexicalizarion pattems 95

distinction here is not one of different causation types. For the latter, an As discussed in note 4, there is no reason to assume that these verbs
increase in participants brings along with it an increment in actional incorporate any re$exive meaning in conjunction with some ‘basically’
content, as in going from the autonomous Z%e snow melred to the other-directed sense. It is quite possible to regard these verbs simply as
agentive John melred rhe snow. Involved here, rather, is a new para- expressing aciions that manifest directly in the actor’s own person. In
meter, one that we will cal1 ‘personation’, pertaining to the role- having such a group of forms, English distinguishes itself from French,
structure that is ascribed to an action. An action complex of certain which must use the reflexive with al1 the corresponding verb forms:
kinds can be taken to manifest either locally, in the body and move-
(52) a. se raser f. se gratter
ments of a single actor (the monadic personation type), or distributive-
ly, with an actor’s body acting on that of a further participant (the b. se laver g. se boutonner
dyadic personation type) . c. se savonner h . s’habiller
A verb root can be lexicalized for just one personation type (either d. se baigner i. se déshabiller
one), taking grammatical augmentation to express the opposite type, or
it can range over both types. Languages exhibit different pattems, with e. . . . (prendre une douche) j. . . . (changer de vêtements)
a bias toward one or another type of lexicalization. Consider, for As already noted, English verbs of the type in (51) generally can also
example, the category of actions involving the use of hands or handled express the dyadic personation type (e.g. I shaved him), and so cover
materials on a body. French, for one language, apparently must the range of lexicalization types. But Atsugewi has a group of verbs like
lexicalize such actions in the dyadic personation type, as actions those in (51) that refer only to the monadic type. To express the dyadic
performed on a diferent person’s body. For the case of action on an type, these verbs must add an idectional element - usually the
actor’s own body, grammatical derivation must be employed - here, the benefactive s&ix -iray. With this set of forms, Atsugewi behaves in a
reflexive: way quite complementary to that of French. One example:
(49) a. Je raserai Jean (53) a. instrumental prefix +
1 will-shave John verb root: -cu-s$ ‘comb the hair
‘1 will shave John’ inf3ectional af&-set: s- ‘- w- 2 ‘I - subject’
b. Je me raserai /s-‘-w-cu-sfial-a/ * [&us$ii]
1 myself will-shave ‘1 combed my hair’
‘1 will shave’ b. instrumental prefix +
English, too, has many verbs with this personation-type, for example: verb root: -CU-S$ ‘comb the hair’
benefactive suffix: -iray ‘for another’
(50) a. 1 cut/bandaged/tickled John
inflectional affix-set: m- w- -isahk ‘1 - subject, thee - object’
b. I cut/bandaged/tickled /m-w-tu-scaf-iray-isahk/ j[mcus@ré&&i]
‘1 combed your hair’
But there is a sizable group of English verbs whose simplest form can Ameritan Sign Language appears to lexicalize exclusively in the
- in addition to indicating different-person referente - also express the monadic personation type for refeting to a certain class of actions,
Agent acting on his own body, thus incorporating the monadic persona- those that in any way involve the torso. Signs for such actions intrinsical-
ti011 type as well: ly represent them as a person would perform them on himself. These
(51) a. I shaved f. I scratched (toa hard)/Don’t scratch! signs must be augmented by additional gestures (such as a shift in
b. 1 washed g. 1 buttoned up body direction) in order to indicate that the actions are performed on
c. 1 soaped up someone else. For example, a woman signer can assert that she had put
h. 1 dressed
on earrings by (among other gestures) bringing her two hands toward
d. I bathed i. 1 undressed her ears. However, to assert that she had put the earrings on her mother
e. 1 showered j. I changed (who has been ‘set up’ at a certain point of nearby space), she cannot

96 LEONARD TALMY Lexicalization patterns 9 7

simply move her hands outward toward where her mother’s ears would Table 2.7. Valance properties for selected English verbs
be. Rather, she only begins by moving her hands outward, but then
shifts her body direciion slightly and adopts a distinci facial expression - (i) Valence properties for emanate, emil, nnd rudiate
Figure as subject Ground as snbject
indicating that her torso is now representing that of her mother - and Light emanates from the sun *The sun emanates light
curves her hands back around, moving them again to her own ears. That ‘Light emits from the sun The snn emits light
is, an additional gestural complex is necessary to indicate that the Light radiates from the sun The sun radiates light
referent action is to be understood as other-directed.” (ii) Valence properties for steal, rob and rip ojf
Figure as direct objeci Ground as direcr object
1 stole his money from him ‘1 stole him of his money
1 . 9 Valeme *I robbed his money from hirn 1 robbed him of his. money
I .g. I General considerations 1 ripped his money off from him 1 ripped him off (? of his money)
In conceptualizing an event that involves severa1 different entities in (iii) Valence pnrrems wifh the Figure exhibtiing a ‘to’-¿ype Futh (F = Figure, G = Ground,
distinct roles, one is able to direct greater attention to some one of these A = Agent)
non-ugenrive ugenfive
entities than to the others or, perhaps, to adopt its actual perspective baric arder Perfume (F) suffused I(A ) sutTused perfume (F) through
point. A secondary degree of attention or perspective-taking, further, tbrough the room (G) the room (G)
can be accorded to some second entity. Such cognitive forms of focusing inverted arder The mom (G) suffused 1 (A) SM the room (G)
witb perfume (F) with perfume (Ff
in are indicated liiguistically by a variety of devices. One device is to
make the focused element the grammatical subject - or, for assigning (iv) Vulence punerns wtih rhe Figure exhibifing a ‘fium’qpe Fafh
non-agentive ugenfive
secondary focus to an additional element, to make that the direct basic arder The blood (F) drained I (A) drained the blood (F) from
object. (Within the scope of our description, it will suffice to adopt from hii veins (G) his veins (G)
simple notions of the grammatical relations ‘subject’ and ‘direct object’, inverted urder His veins (G) drained of 1 (A) drained his veins (G) of
their blood (F) their blood (F)
and to associate these with the case markings ‘nominative’ and ‘accusa-
tive’ in the languages that have these.) Now, a lexical verb that refers to (The Word sIow¡y can be inserted in the preceding sentences for smoother reading.)
a multi-roled event can have built-in constraints on its freedom to assign
focus. It can be limited to taking only a particular one of the element with radiare, which accommodates either. Thus, emanate incorporates
types as subject (or diiect object), and so lexicalizes focus on that focus on the Figure (the radiation) and emir does this for the Ground
element type. In other instances, a single verb can accommodate (the radiator), while radiare can incorporate either focus.
different element types in the focus position, and so has a range of We can demonstrate a similar relationship with an agentive example.
Iexicalizations. Such focusing properties are here called the ‘valence’ of Sreal, rob, and rip ofa refer to the same event and take nominals for
a verb. Traditionally, the rerm valence has been used to refer (either the Agent, Figure, and Ground roles.% Allgive the Agent primary focus
solely or additionally) to the number of distinct element types occurring as subject. But for secondary focus as direct object, sreal selects the
in association with a verb. In this chapter, the issue of element number Figure (the possessions) while rob selects the Ground (the possessor).
arises only in the treatment of causation and personation. Valence here Rip off accommodates either.
is used just for the particular surface case assignment(s) that a verb Some verbs - suffuse and drain are examples - can accommodate their
exhibits, given a fured number of certain types of elements in association nominals in either the basic Figure-before-Ground arder or the inverted
with it. Ground-before-Figure order in both the non-agentive and the agentive.
The notion of incorporated valence can be effectively demonstrated Under inversion, the Figure acquires one of two ‘demotion particles’ -
where there are two verbs whose subject limitations together equal the of when it exhibits an underlying ‘from’-type Path, as with druin, and
range of subject possibilities of a third verb. This is the case with with for other Path types, as with sume (some languages use different
emmate and en& on the one hand and radiare on the other. Al1 three of cases for this). Thus, the full array of these two verbs’ forms in effect
these verbs refer to roughly the same event, an event having both a constitutes a parad@ against which other verbs, more liited in one
Figure element and a Ground element. But emanare requires @e Figure respect or another, can be compared. See Table 2.7 for the valence
as subject, while emit requires the Ground as subject - as contrasted properties of al1 the preceding English verbs.35

LEONARD TALMY Lexicalization patterns 99
98

In the same way as with aspect and causation, a language can have Table 2.8. Derivatimal pattems for a#ect verbs focused on the Stimulus
grammatical devices for use with a verb of one valence type in order to or the Experiencer
express a different type. German has this arrangement for cases of the
Stimuhs as subject * fiperieruer us subject
secondary focus sort. Its prefix be- can indicate a shift in secondary focus It frightens me 1 am frightened of it
from the Figure onto the Ground: It pIeases me 1 am pleased with it
It interests me I am interested in it
(54) a. Ich raubte ihm seine Tasche
Experience as subject * Stimulus as subjeci
1 stole ~~I~(DAT) his(Acc) Wallet 1 fear it It is fearfui to me
‘1 stole his wallet from him’ (Figure as direct object) 1 like it It is likeabb to me
b. Ich beraubte ihn s e i n e r Tasche 1 toathe it It is Ioathsome to me
1 sHlFr-stole hh(ACC) hiS(GEN) wallet
‘1 robbed him of his wallet’ (Ground as direct object)36
Where a language, as here, has a grammatical device for getting to a 1.9~2 Valence in verbs of affecf
particular valence type, it might tend to have relatively few verb roots Consider verbs of affec? with respect to valence. These verbs generally
Iexicalized in that type. In fact German appears to have fewer verb roots require either the Stimulus or the Experiencer of an affective event as
like our ro& and pelt, roots that intrinsically take the Ground as direct the subject. Accordingly, they incorporate focus on either the qualities
object, using instead its complexes of Figure-taking root plus valence- of the Stimulus or the state of the Experiencer. Compare this lexicaliza-
shifter, like be-raub(en) and be-werf(en). The two languages contra9 in tion difference in frighren and fear, which refer to roughly the same
a similar way in what can be called verbs of giving, this time as to how affective situation:
they indicate focus on (and, hence, the point of view of) the giver or the W a. That frightens me - Stimulus as subject
receiver. Both languages do have cases where the distinction is indicated
b. I fear that - Experiencer as subjed7
by distinct verb roots of complementary valence type:
155) give teach For verbs Iexicalized in either valence type, there are grammatical, or
get (in the sense of ‘receive’) leam grammatical-derivational, means for getting to the opposite type. Thus,
geben lehren a verb with a Stimulus subject can generally be placed in the construc-
tion ‘be - V + PP - Prep’ (not a passive: the preposition can be other
kriegen lemen
words than by) to bring the Experiencer into subject position. And a
But in other cases, English has two verb roots where German has only verb with an Experiencer subject can often figure in the construction ‘be
one, one lexicalized with focus on the receiver. A pretil ver- reverses - V + Adj - to’, which places the Stimulus as subject. See Table 2.8.
the perspective to the giver’s point of view: While possibly al1 languages have some verbs of each valence type,
(56) sebo bequeath lend I they differ as to which type predominates. In this respect , English seems
buy inherit borrow to favor lexicalizing the Stimulus as subjecL3* While some of its most
verkaufen vererben verleihen verborgen colloquial verbs (Me, wanr) have the Experiencer as subject, the buIk of
kaufen erben leihen borgen its vocabulary items for afEect focus on the Stimulus, as we see in Table
2.9.39
This is illustrated in (57). By contrast with English, Atsugewi roots appear to have Experiencer
(57) a. Ich kaufte das Haus von ihm subjects almost exclusiveiy. Virtually every afEect-expressing verb (as
I boughtthe house from him well as adjective in construction with ‘be’) elicited in fieldwork was
‘1 bought the house from him’ lexicalized with an Experiencer subject. To express a Stimulus subject,
b. Er verkaufte mir das Haus these forms took the suffix -ah&. For one example see Table 2.10.~
he bought(REvERsE) me(DAT) the house It may be that the boundaries of the ‘affect’ category here are toa
‘He sold me the house’ encompassive or misdrawn for good comparative assessments. There

100 LEONARD TALMY Lkzxicalizadon patterns 101

Table 2.9. Afect verbs in Engltih Table 2. II. ‘Coenitive’ verbs in English

Stirnulus as subject Stimul~~ as snbject Experiencer as subject
strike ocar to k n o w tbink consider remember leam
please key up il.SIOlliSh incense WOrrY seem to dawn on real& feel suspect forget discover
satisfy tu-n on awe bothér infuriate concem remind . . of believe d o u b t imagine wonder about find out
gratify interest wow irk outrage trouble
ClXllf0t-t engage confuse bw nliff diStESS
soothe captivate pme vex pu1 out upaet
calm intrigue perplex pique disgruntle diStWb may be smaller categories following more ‘natural’ divisions that reveal
charm fascinate ws* peeve frustrate disconcert i,:
beguile baffíe nettle chagrin unsettle
more about semantic organization. For example, a ‘desiderative’ cate-
Ere: entrance bewilder initate embarrass shake up gory might well be separated out by itself: al1 the English verbs of
ticlde bewitch @de provoke abash discombobulate ‘wanting’ listed in Table 2.9 have Experiencer subjects, and this
delight tantalti swefy pau cow frighten arrangement might be universal. Thus, although colloquial expressions
thrill matter to dumbfound aggravate shame scare
transport bore flabbergast grate on humiliate ahrm with the opposite valence occur in other languages:
move surprise shock piss 08 disgust grieve
stir startle dismay exasperate gross out hurt (59) a. Yiddish:
arouse arnaze VPd anger revolt pain Mir vilt xix esn
excite astound honify rile torment me-ro wants REFL to eat
Experiencer as subject b. Samoan:
like marvel over WXH tust for abhor worry about ‘Ua sau (liate a’u) le fia ‘ia
enjoy wonder at feel lilce crave deplore grieve over
care for trllst desire need anger over sorrow over AsP come (to me) the want (to) eat
groove on respect prefer covet fume over regret ‘A desire for eating has come on me (1 feel like eating)’
f=Y esteern wish for e=vy seethe over me
reIish admire hope for dislike gloat over hurt from they are derived constructions based on verb roots with Experiencer
love appreciate hanker after resent distrust ache from subjects. (However, Kaluli of New Guinea may possibly be a language
adore value hunger for hate fear suffer fmm
delight in Ihirst for detest dread bar in which all mental verbs - including those of ‘wanting’ and ‘knowing’ -
Ihrill to :%h long for despise stand put the Experiencer in the surface case that identifies it as the affected
exult over revere yearn for loathe tolerate argument (Bambi Schieffelin, personal communication).) Perhaps, too,
one should separate out an ‘assessment’ category for notions like
Table 2.10. Derivation uf Experiencer-subject verb rooti tu Stimulus- ‘esteem’, ‘value’, ‘prixe’; in Table 2.9 the English verbs for these notions
subject in Atsugewi again al1 require Experiencer subjects. We had already separated out a
Experiencer us subject
‘cognitive’ category for the more intellective mental processes. Verbs of
verb root: -lay- ‘to consider as good’ this category were excluded from the affect list above, and again English
instrumental prefixz sa- ‘by vision’ seems to favor Experiencer as subject for them, as shown in Table 2. I I.
derivational suñtx: -im (no specitíc meaning: occurs here A single semantic-cognitive principie might account for all these
idiomatically)
‘1 - subject, 3d person object’ correlations between category of mental event and lexicalization
tendency: subjecthood, perhaps because of its frequent association with
agency, may tend to confer upon any semantic category expressed in it
Derived to: Srim&s as srrbject some initiatory or instigative characteristics. Accordingly, with Stimulus
verb root: -lay- ‘to consider as good’
instrumental prefix: sa- ‘by vision’ as subject, an extemal object or event (the stimulus) may be felt to act
valence-shifting suffix: -ah& ‘fmm Stirnutus to Experiencer’ on an Experiencer so as to engender within him/her a particulal .?ental
tiectional affix-set: ‘_ w- *a ‘3d person s u b j e c t ’ event. Conversely, with Experiencer as subject, the mental event may
‘lt ~/~e;;;~-ah&~/ j [&al.ayáh&a]
be felt to arise autonomously aÍrd to direct itself outward toward a
selected object. For example, a mental event of ‘wanting’ might be

i‘

psychologicahy experienced across cultures as a self-originating event, Table 2.12. Sateliites as verb prefìxes in Geman, Latin, and Russian
and so, by this principie, have a preponderant tendency across lan-
guages to correlate with Experiencer subjecthood. a. German
kparabk prq5.x ‘ i n s e p a r a b l e ’ prq5x
satellite: +entzwei +5er-
2.0 Satdlites verb complex: brechen +entzwei (entzweibrechen) brechen +zer- (zerbrechen)
ex. sentence: Der Tisch brach entzwei Der Tiich zerbrach
Here we wilI examine the representation of certain semantic categories ‘The tabie broke in two’ ‘The table broke to pieces’
by a type of surface constitnent that has not been generahy recognized b. Lah: c. Russian:
as such in the linguistic literature, one that we terna a ‘satelhte’. Present PM=
sateliite: +in- l Y-
in many if not al1 languages, satellites are certain immediate constituents verb complex: volare +in- (mvolare) letet’ l v- (vletet’)
of a verb root other than inflections, auxiliaries, or nominal ex. sentence: Avis involavit Ptica vletela
arguments . 41 They relate to the verb root as periphery (or modifiers) to ‘The bird flew in’ ‘The bird Aew in
a head. A verb root together with its satellites forms a constituent in its
own right, the ‘verb complex’, also not generally recognized. It is this
constituent as a whole that relates to such other constituents as an Tibeto-Burman languages. In the case of Lahu, Matisoff (1973) has
inflectional &-set, an auxiliary, or a direct object noun phrase. In called any such word a ‘versatile verb’. A third example is any of the
some cases, elements that are encountered acting as satellites to a verb non-inflectional aties on the verb root in the Atsugewi ‘polysynthetic
root otherwise belong to particular recognizable grammatical verb’ .42 We now examine a range of types of semantic material that
categories; therefore, it seems better to consider the satellite role not as appear in satellites.
a grammatical category in its own right but as a new kind of grammatical
relation . 2.1 Path
The satellite is easily illustrated in English. It can take the form of
The satellites in English are mostly involved in the expression of Path.
either a free word or an affix (satellites are marked here by the symbol f Generally, the Path is expressed fully by the combination of a satellite
that, in effect, ‘points’ from the satelhte to its head, the verb root):
and a preposition, as in (62a). Rut usually the satellite can also appear
(60) satellite: verb complex: example sentence: alone, as in (62b). The ellipsis of the prepositional phrase here generally
fover start *ver The record started over requires that its nominal be either a deictic or an anaphoric pronoun
flT¡iS- fire fmis- The engine misfired (i.e., that the Ground object be uniquely identifiable by the hearer):43

(62) a. I ran OUT of the house
As many as four such satellites can appear together in a verb complex:
b. (After rifling through the house,) 1 ran out [Le., , . . of it]
(61) Come fright fback fdown fout from up in there!
(said, for example, by a parent to a Child in a treehouse) Some symbolism here can help represent the semantic and gramma-
tical situation. The symbol > is placed after a preposition, in effect
The term traditionally applied to the above element in English is ‘verb pointing toward its nominal head. Thus this symbol together with f
particle’ (see Fraser 1976). The term ‘satellite’ has been introduced in enclose the ful1 surface expression (the satellite plus preposition) that
order to capture the commonality between such particles and compar- specifies Path, as illustrated in (63a). For a still finer representation,
able foims in other languages. Within Indo-European, such forms parentheses are used to mark off the portion that can be optionally
include the ‘separable’ and ‘inseparable’ prefixes of German and the omitted, and F and G indicate the locations of the nominals that
verb prefixes of Latin and Russian as shown in Table 2.12. function as Figure and Ground,. as shown in (63b):
Another kind of satellite is the second element of a verb compound in
Chinese, called by some the ‘resultative complement’. Anothef example (63) a. fout of>
is any non-head word in the lengthy verbal sequences typical of b . F . . . fout (of> G)

ro4 LEONARD TALMY Lexicalization patterm 105

English has quite a few fath satellites. Some are presented in the (67) a. Ya vbeial (v dom)
sentences below, here without any final Ground-containing phrase: I in ran (into house(Acc))
I ran in (-to the house)
(64) Path satellites in English
b. Ya vybeial (ix doma)
1 ran in It Aew up He ran along
1 OUt rall (OUt Of houSe(GEN))
1 ran 0r4t It flew down He ran around
I ran out (of the house)
1 got un 1 went above He ran pa.sr/by
1 got 08 1 went below He ran away
She carne over He ran through He ran back We want to emphasize for al1 these Path examples that satellites
It toppled over He ran across She carne forth should be well distinguished from prepositions. No confusion can occur
They roIled apart in most Indo-European languages, where the two forms have quite
They slammed together distinct positional and grammatical characteristics. For example, in
Latin, Classical Greek, and Russian (6. (66) and (ti)), the satellite is
In addition, English has a number of Path satellites that would not be bound prefrxally to the verb while the preposition accompanies the noun
generahy recognized as such, i.e., as being in the same semantic (wherever it turns up in the sentence) and govems its case. Even where
category as those of (64): a satellite and a preposition with the same phonetic shape are both used
together in a sentence to express a particular Path notion - as obten
(651 More Path satellites in English happens in Latin, Greek, and Russian (again, see (66) and (67)) - the
F . . . floose (from> G) The bone pulled loose (from its socket) two occurrences are still formally distinct. However, a problem arises
F . . . ffree (from> G) The coin melted free (from the ice) for English which, perhaps done among Indo-European languages, has
F . . . fdear (of> G ) She swam clear (of the oncoming ship) come to regularly position satellite and preposition next to each other in
F . . . fstuck (to> G) The twig froze stuck (to the window) a sentence. For some of these juxtapositions, a kind of merged form has
F . . . +fast (to> G ) The glaze baked fast (to the clay) developed, while for others - especially where two occurrences of the
F . . . fun- (from> G) The bolt must have unscrewed (from same shape might be expected - one of the forms has dropped (we treat
the plate) this next). Nevertheless, there are stiI1 ways in which the two kinds of
F . . . *ver- d> G The eaves of the roof overhung the forms - satellites and prepositions - distinguish themselves. First, it is
garden _ only a preposition that will disappear when the Ground nominal is
F . . . +under-4 > G Gold leaf underlay the enamel omitted: a satelhte remains. Next, the two classes of forms do not have
G . . . +full (of> F ) The tub quickly poured full (of hot identical memberships: there are forrns with only one function or the
water) other. For example, together, upar?, and f’rth are sateilites that never
act as prepositions, while from, at, and toward are prepositions that
The languages in most branches of Indo-European have Path systems never act as satellites. Furthermore, forms serving in both functions
that are homologous with the one just seen for English. That is, tbey often have different senses in each. Thus, ro as a preposition (‘I went to
also use a satellite and a preposition, with the prepositional phrase the store’) is different from CO as a satellite (‘1 carne to’), and satellite
generally omissible. Tbis is illustrated here for Russian (see Talmy 1975 over in its sense of ‘rotation around a horizontal axis’ (‘It
for an extensive treatment of such forms in this language): fell/toppled/tumed/flipped over’) does not have a close semantic
counterpart in prepositional over with its ‘above’ or ‘covering’ senses
(66) Path expressions in Russian
(‘over the treetop’ , ‘over the Wall’).
fv- V + ACC> ‘into’ fpere- &rez + Acc> ‘across’ We look more closely now at the special feature of the English Path
+~y- iz + GW> ‘out of’ fpod- pod + ACC> ‘(to) under’
fna- na + Acc> ‘onto’ +pod- k + DAT> ‘up to’ system; it is Worth going into because the same feature will appear again
fs- S + GEN> ‘ o f f of fpli- k + DAT> ‘ir&-arrivai-at’ in the Mandarin system that we treat next. English has a number of
fdo- do + GEN> ‘al1 the way to’ forms like past that behave Iike ordinary satelliteswhen there is no fmal

ro6 LEONARD T A L M Y Lexicalization pattem 107

nominal, as in (68a), but appear without any preposition when there is a These satellites participate in Path expressions of either the coalesced or
6nal nominal, as in (68b): the uncoalesced type. The only apparent diñerence from English is an
order distinction: the object of the coalesced fonn follows the verb
a. (1 saw him on the comer but) I just drove past complex, whereas the prepositional phrase of the uncoalesced form
b. 1 drove past him precedes it (as is general with prepositional phrases of any kind). Some
satellites can participate in both constructions. One of these is the
A form like that in (68b) has properties of both a satellite and a satellite meaning ‘past’, which we see here in two different sentences
preposition. It receives the heavy stress of a satellite (a preposition that receive the same translation in English:
receives Iight stress, as in ‘1 went to him’). But Iike a preposition it is
always positioned before the nominal (an ordinary satellite may follow a (72) F. . . +guò (-$> G -biàn) (coalescence of satellite and
direct object noun, and must follow a pronoun, as in ‘1 drove him in’). past side preposition)
Tbe different English accentual and positional types can be contrasted Píng-zi pião guò shi-t6u páng-biãn
thus: bottle float past rock(‘s) side
‘The bottle floated past the rock’
(69) a. Iwenttõhim - with a preposition alone
b. 1 followed him ín - with a satellite alone
(73) F- . . euò (cóng> G -biãn) (the uncoalesced form with both
c. 1 went ín t6 him (where he sat)- with both a satellite and a past from side a satellite and a preposition)
preposition Píng-zi cóng shí-tóu páng-biãn pião guò
d. 1 went pást him - with a satellite-preposition bottle from rock(?) side float past
‘The bottle Aoated past the rock
Because of its special behavior, a form Iike past might be considered a
coalesced version of a satellite plus a preposition - a satellite-preposi- 2.2 Path + Ground
tion @st treated in Talmy 1972) - as suggested symbolically in (70a). In a conflation pattem distinct from the preceding one, a satellite can
Or, as is assumed here, it can be considere! a real satellite that happens express at once both a particular Path and the kind of object acting as
to be coupled with a zero preposition, as suggested in (70b): Ground for the Path. Satellites of this sort seem to be rare in the
languages of the world. However, they constitute a major type in certain
(70) a. +past> b. F. . . +past ($> G) Amerindian Ianguages. English does have a few examples, which can
serve to introduce the type. One is the form heme in its use as a satellite,
Mandarin Chinese has Path satellites and constructions that are where it has the meaning ‘to his/her/. . , home’. Another is the form
entirely homologous with those of English. A number of these satellites shut, also in its satellite use, where it means ‘to (a position) across an
are listed here (they varioudy may, cannot, or must be further followed opening’. These forms are here illustrated in sentences, optionally
by the satellite for ‘hither’ or for ‘thither’): followed by prepositional phrases that amplify the meanings already
present in the satellites:
(71) tsu :;Fr; tguò ‘across/past’
+lái +d klp off (74) a. She drove home (to her cottage in the suburbs)
fshàng ‘up’ +diào ‘off (He ran ofl’ b. The gate swung shti (across the entryway)
+xià ‘down’ +2h ‘away’
+fm ‘in’ fhui ‘back Atsugewi is one language which has such satellites as a major
+chü ‘out’ flõng ‘together’ system.44 It has some f~fty forms of this sort. We can illustrate the system
fdào ‘al1 the way (to)’ fkãi ‘apart/free’ by listing the fourteen or so separate satellites that together are roughly
420 ‘atopple (Le., pivotally over)’ +sàn ‘ascatter’ equivalent to the English use of into with different particular nominals.

ro8 LEOKARD TALMY Lexicdization pactems 109

[A ‘+’ here indicates that the satellite must be followed by one of b. verb root: -lup- ‘for a small shiny spherical
-irn/-&, ‘hither’/‘thither’): object to move/be Iocated’
Path + Ground sateilites in Atsugewi directional suffix: -mik- ‘into the face/eye(s) of
(75) _
-Et ‘into a liquid someone’
-cis ‘into a fire’ instrumental prefrx: phu- ‘from the mouth - working
lisp -u- + ‘into an aggregate’ (e.g. bushes, a crowd, a rib-cage) egressively - acting on (the
-wani ‘down into a gravitic container’ (e.g. a basket, a Figure)
cupped hand, a pocket, a lake basin) inflectional affix-set: m- w- -a ‘thou - subject , 3d person
-wamm ‘into an areal enclosure’ (e.g. a corral, a lieId, the object (factual mood)’
area occupied by a pool of water) /m-w-phu-lup-mik--“/ 3 [mphol-úphmik-a]
-ipsn” + ‘(horizontally) into a volume enclosure’ (e.g. a Literal: ‘You caused it that a small shiny spherical object
house, an oven, a crevice, a deer’s stomach) move into his face by acting on it with your mouth working
-tip -u* + ‘down into a (Iarge) volume enclosure in the ground egressively’
(e-g. a cellar, a deer-trapping pit) Instantiated: ‘You spat your candy-hall into his face’
-ikn + ‘over-the-rim into a volume enclosure’ (e.g. a
gopher hale, a mouth)
-ikc ‘into a passageway so as to cause blockage’ (e.g. in 2.3 Patient: (Figure/) Ground
choking, shutting, walling off) Another type of satehite is one that indicates the Patient of an event
-iks” + ‘into a comer’ (e.g. a room comer, the Wall-floor being referred to. Though apparentiy rare otherwise, such sateIlites do
edad constitute a major system in some Amerindian languages, those known
-mika ‘into the face/eye (or onto the head) of someone’ as ‘noun-incorporating’. These languages inciude an afJ?xaI form of the
-miE ‘down into (or onto) the ground satellite within their polysynthetic verb. Caddo is a case in point. Here,
-c-k” + ‘down into (or onto) an object above the ground the satellite gives a typicahy more generic identification of the Patient.
(e.g. the top of a tree stump) The sentence may also contain an independent nominal that @ves a
-iiS !horizontally into (or onto) an object above the typicaliy more specific identification of the same Patient, but the
ground’ (e.g. the side of a tree trunk) satellite must be present in any case. Here first are some non-motion
Instances of the use of this satellite system can be seen in the Atsugewi examples, with (77a) showing the Patient as subject in a non-agentive
sentence, and (b) and (c) showing it as direct object in agentive
examples appearing earlier, (Tga, b, c), (4oa, b), and (48); two further
exampies are given in (76). sentences:
(77) a. Qiku3 hák-n&&-ni-kah-sa3j [Xniku? háhnisánkáhsap]
(76) a. uerb root: -sia+ ‘for runny icky material to church PRoG-house-bum-PROG
move/be located’ LiteraUy: ‘The church is house-burning
directional suffix: -ipsn” ‘into a volume enclosure’ (i .e . , building-buming)’
deictic suEx: -ik* ‘hither’ Loosely: ‘The church is burning’
instrumental prefix: ma- ‘from a person’s foot/feet
acting on (the Figure)’ b. cú-cu’ /can-yi-dapk-ah j [cúcu? kannida’kah]
inflectional aEix-set: ‘- w- -a ‘3d person subject (factual milk hquid-find-PAST
mood)’ LiteraUy: ‘He liquid-found the milk
/ ‘-w-ma-staq-ipsn”-ik*-a/ j[Aasia$psnuk*a] Loosely: ‘He found the milk
Literal: ‘He caused it that runny icky material move hither c. widil dú ‘32~yi-dapk-ah $ [widiS dânnidapkah]
into a volume enclosure by acting on it with his feet’ salt powder-find-PAST
Instantiated: ‘He tracked up the house (coming in with LiteralIy: ‘He powder found the Salt’
muddy feet)’ Loosely: ‘He found the Salt’

II0 LEONABD TALMY Lexicalization patterns III

Without the independent noun, the last example wouId work in this (82) Nez Perce Manner prefixes
way: ?psqi- ‘walking’
(78) dá%-yi-da?k-ah ‘He powder-found it’/‘He found it ‘something wilém- ‘running’
powdery)’ wat- ‘wading’
siwi- ‘swimming-on-surface’
ln Caddo’s general pattem for expressing Motion, the verb root tukwe- ‘swimming-within-Quid
indicates fact-of-Motion together with Path, in the manner of Spanish. we- ‘flying’
The incorporated noun can under limited conditions - it is not yet dear tuke- ‘using a cane’
what these are - indicate the Figure, as in this locative example: ceptukte- ‘crawling’
nti&ya-%& 3 [dahkhih tisáy’ah] tukweme ‘(snake) slithering’
(79) yak*aMh
woods edge-uc house-be-ms Wu-I- ‘(animal) walking/(human) riding (animal at a
Literally: ‘At woods edge it-house-is’ waik)’
Loosely: ‘The house is at-the edge of the woods’ ‘(animal) galloping/(haman) galloping (on animal)’
‘(heavier object) floating-by-updraft/wafting/
Usually, the incorporated noun indicates the Ground: gliding’
3iyé.- ‘(lighter object) floating-by-buoyancy’
WI a. wákas na-yawat-ya-ynik-ah j [wáekas táywacáynikah] wis- ‘travelling with one’s belongings’
catie PL-water-enter-PAsT kipi- ‘tracking’
Literally: ‘Cattle water-entered ti&ek- ‘pursuing (someone: or3J)’
Loosely: ‘The cattle went into the water’ cú-- ‘(plural@) in single frle’
b. nisah-nt-káy-watak-ah + [tisáncáywakkah] til- ‘on the warpath/to fight’
house-penetrateltraverse-PAsr qisim- ‘in anger’
Literally: ‘He-house-traversed
Loosely: ‘He went through the house’ Assuming that polysynthetic forms arise through boundary and sound
changes among concatenated words, one can imagine how a Nez
Perce-type system could have developed from a Spanish type. OriginaLly
2.4 Manner independent words referring to Manner came regularly to stand next to
Another uncommon satellite type is one expressing Manner. An the verb and then became affixal (and in most cases also lost their usage
extensive system of such satellites is found in Nez Perce, another elsewhere in the sentence). Indeed, one can imagine how Spanish might
polysynthetic language of North America (see Aoki 1970). In Motion evolve in the direction of Nez Perce. The preferred position for
sentences, the verb root in this language is Iike that of Spanish: it Manner-expressing gerunds in Spanish is already one adjacent to the
expresses Motion + Path. But at the same time, a prefix adjoining the verb, as in:
root specifies the particular Mamier in which the Motion is executed.
An example of this arrangement is given in (SI). (831 Entró corriendo/volando/nadando/ . . . a la cueva
he entered running flying swimming to the cave
(81) /hi-quqúm-láhsa-e/ $J [hiqqoláhsaya]
3d person-galloping-goup-PAST One could imagine the few changes that would be necessary to take this
Literally: ‘He/she ascended galIoping’ into the Nez Perce system.
Loosely: ‘He galloped uphill
2.5 Cause
We list here a selection of Nez Perce Manner prefixes. Note that not just A kind of satellite found in a number of languages, at least in the
locomotive manners are expressed, but also ones of affect (‘in anger’) Americas, has traditionally been described as expressing ‘Instrument’.
and activity (‘on the warpath’): However, these forms seem more to express the whole of a Cause event.

II2 LEONARD TALMY Lexiculization pattem 113

This is because, at Ieast in the familiar cases, not only the kind of fuh- ‘from a linear object acting circumpivotally (swinging)
instrumental object that is involved is indicated, but also the way in on P’ (as in pounding, chopping, batting)
which this object has acted on a Patient (to cause an effect). That is, a fra- a. ‘from a linear objec? acting obliquely on P ’ (as in
satellite of this sort is equivalent to a whole subordinate clause digging, sewing, poling, leaning)
expressing causation in Engfsh. In particular, a satellite occurring in a b. ‘from a linear/planar object acting Iaterally along
non-agentive verb complex is equivalent to a fiom-clause, as in (to take the snrface of P’ (as in raking, sweeping, scraping,
an actual example in translation) : ‘The sack burst from a long thai object plowing, whittling, smoothing, vising)
poking endwise into if. And the same satellite occurring in an agentive fta- ‘from a linear object acting within a liquid P’ (as in
verb complex is equivalent to a by-clause, as in ‘I burst the sack by stirring, paddling)
poking a long thin object endwise inro it’. fka- ‘from a linear object moving rotationally into P’ (as in
Perhaps the greatest elaboration of this satellite type occurs in the boring)
Hokan languages of northem California, with Atsugewi having some +Illi- ‘from a knife cutting into P’
two dozen forms (see Talmy rg72:84-Ig5, 407-67). Here, most verb l -u- ‘from a (flexible) linear object pulling on or inward
roots must take one or another of the Cause satellites, so that there is upon P’ (as in dragging, suspending; girding, binding)
obligatory indication of the cause of the action expressed by the verb body parts in aciion
root (some verb roots cannot take these satellites, but they are in the fb- ‘from the hand(s) - moving centripetally - acting on P’
minority). The ful1 set of these satellites subdivides the semantic domain (as in choking, pinching)
of possible causes fairly exhaustively. That is, any perceived or con- &- ‘from the hand(s) moving manipulatively - acting
ceived causal condition will likely be covered by one or another of the on p’
satellites. The majority of the Atsugewi Cause satellites, those in fma- ‘from the foot/feet acting on P’
commonest use, are listed below. They are grouped here according to +ti- ‘from the buttocks acting on P’
the kind of instrumentality that they specify. As in other Hokan *wi- ‘from the teeth aciing on f’
languages, they appear as short preíixes immediately preceding the verb l pi- ‘kom the mouth - working ingressively - acting on P’
root: (as in sucking, swallowing)
+phu- ‘from the mouth - working egressively - acting on p’
(84) Atsugewi Cause satellites (P = the Patient, E = the Experiencer) (as in spitting, blowing)
nmalforces +pu- ‘from the lips acting on P’
fea- ‘from the wind blowing on P’ fhi- ‘from any other body part (e.g. head, shoulder) or the
+=- ‘from Aowing liquid acting on P’ (e.g. a river on a whole body aciing on P ’
bank) sensations
fka- ‘from the rain acting on P ’ ‘from the visual aspect of an object acting on E’
fsa-
*ra- ‘from a substance exerting steady pressure on P ’ (e.g.
*ka- ‘from the auditory aspect of an object acting on E’
gas in the stomach) *tu- ‘from the feel of an object acting on E’
+uh- ‘from the weight of a substance bearing down on P’ ‘from the tastelsmell of an object acting on E’
+p+
(e.g. snow on a limb)
‘from “gravity” (the tendency of things to fall) acling Instances of these satellites in use in a verb have appeared in examples
on p’ (Iga, b, and c) and (76a and b), to which the reader is referred.
fmiw- ‘from heat/fire acting on P’
2.6 Motion-relateti satellites extenhtg the motion fypoiogy
objecís in action Table 2.2 (section 1-4) showed the three major categories into which
fcu- ‘from a linear object acting axially on
P’ (as @ poking, languages fall in their treatment of Motion. The typology was based on
prodding, pool-cueing , piercing , propping) which element of a Motion event is characteristically expressed in the

rr4 LEONARD TALMY Lexicalization patterns 115

Table 2.13. Typology of motion verbs and their satellites wn ‘contiriue Ving without stopping’
We talked/worked on into the night
The particular components of a motion event ‘resume where one had left off in Ving’
Language/Ianguage family characterktically represented in the: She stopped at the gas station first, and then she drove on from
Verb root Satellite there
‘go ahead and V against opposition’
Romance
Semitic b
He was asked to stay on the other side of the door but,
FOlJQESkUl Motion + Path adamant, he barged on in
Ne2 Perce )Manner faway ‘continue Ving (with dedication/abandon)’
Caddo )(Figure/)Ground [Patient] They worked away on their papers
1
They gossiped away about all their neighbors
Indo-Europeau (all?) ‘feel free to embark on and continue Ving’
except Romance Path
chinese
‘Would you like me to read you some of my poetry?’ ‘Read
away !’
Atsugewi (most northern Motion + Figure a. Path + Ground fdong ‘proceed in the process of Ving’
FIokan) b. Cause We were talking along about our work when the door suddenly
burst open
foff ‘V al1 in sequence/progressively’
I read/checked off the names on the list
verb root (with ‘fact of Motion’, which always appears there). For each Al1 the koalas in this area have died off
such language type, the next issue is where the remaining elements of ‘V al1 the way into a different (a non-integral/
+uP
the motion event are located. The satellite is the most diagnostic surface denatured) state’
element to Iook at after the verb, and so we can make a revealing The log bumed up in 2 hours
subcategorization by seeing which motion elements characteristically (cp. The log bumed for I hour before I put it out)
appear in the sateIlites that accompany the verbs; see Table 2.13. The dog chewed the mat up in 20 minutes
(cp. The dog chewed on the mat for IO minutes before I took
2.7 Aspect it away)
Many languages have satellites that express aspect. Frequently, these +back ‘V in reciprocation for being Ved
satellites do not indicate purely ‘the distribution pattern of action He had teased her, so she teased him back
through time’ (as aspect was charactetied earlier). This purer form is
mixed with, or shades off into, indications of manner, quantity, inten- Other languages have forms comparable to those of English, though
tion, and other factors. Accordingly, a liberal interpretation is given to often with different, or more varied meanings. Russian is a case in point.
aspect in the examples below.. In this way, we can present together many In addition to several forms Iike those in the English list, Russian has (at
of the forms that seem to be treated by a language as belonging to the least) the following (some of the examples are from Wolkonsky and
same group. The demonstration can begin with English. Though this Poltoratzky, 1961):
language is not usualIy thought of as expressing aspect in its satellites WI Russian aspect satellites
(as, say, Russian is), it is in fact a fully adequate example: ‘V for a while’
+Po-
Ya pogul’al ‘I strolled about for a while’
035) EngIish aspect satellites (V = to do the action of the verb) XoEets’a poletat’ na samolete
fre-/*ver ‘V again/anew’ ‘I’d like to fly for a while on a plane (Le., take a short flight)’
W h e n i t g o t t o t h e e n d , t h e record automatically *ere- ‘V every now and then’
restarted/started over from the beginning Perepada’ut doidi ‘Rains fall (It rains) every now and then’

II6 LEONARD TALMY Lexicalizatin pattems 117

Table 2.14. Atsugewi aspect satellites’ meanings To illustrate the second satellite type:

V’s action is related to: 037) verb root: w- ‘fOF contained solid
material to move/be-
the general temporal flow an ongoing locomotory event located
almost v goandV secondary aspect suñix: -ikc Yo a position blocking
stillv go Ving along p a s s a g e ’ , hence: ‘in
V repeatedly come Ving. along
v in passing going to meet (and give
V again/back, reV
start Vmg V going along with someone to> someone approach-
finkh Ving V coming along with someone inflectional a%x-set: s- ‘- w- 3 ing’ 1 - sbj (3d person -
V as a norm V in following along after someone
V in going to meet someone object) [factual]’
V awhile/stay awhile and V
Vinahurry/hurryupandV independent noun: taki- ‘acoms’
V a little bit/spottily/cutely nominal marker: C
/s-‘-w-acp-ikc-a c tal&/ * [s;acpfk%a c ta&]
Literally: ‘1 caused it that contained solid material - namely,
acoms - move, in going to meet (and give it to) someone
e-a- ‘start viig’ approaching’
Kapli doid’a zapadali odna za drugoy Loosely: ‘I carried out the basket ful1 of acoms to meet him
‘Drops of rain began to fa11 one after another’ with, as he approachedti
fraz- +s’a ‘burst out Ving’
Ona rasplakalas’ ‘She burst out crying’
+pro-/*ere-/ . . . ‘complete the process of Ving’ 2.8 Valence
Pivo perebrodilo ‘The beer has finished fermenting’ In section 1.9 we saw satellites (German be- and ver-, Atsugewi -ah&)
*Po-/ . . . ‘V as one complete act’ involved solely with valence: they signaled shifts for the incorporated
On yeyo pocelova1 ‘He kissed her’ (VS.: was kissing, kept valence requirements of verb roots. There are also satellites that
kissing, used to kiss) basically refer to other notions, such as Path, but themselves incorpo-
fna- +s’a ‘V to satiation’ rate valence requirements. When these are used with verbs that have no
On nayels’a ‘He ate his fill competing requirements, it is they that determine the grammatical
‘V and de-V as one complete cycle’ [only with motion relations of the surrounding nominals. We look at this situation now.
e-
verbsj Consider these Path satellites (or satellite + preposition combina-
Ya sletal v odin mig na poku tions) referring to surfaces:
Lit.: ‘I there-and-back-flew in one moment to post-Office’ 038) a. Water poured unto the table
‘1 got to the post o@ce and back in no time’ - ‘to a point on the surface of
b. Water poured al1 over the table
Within its affixal verb complex, Atsugewi has certain locations for a - ‘to al1 points on the surface of’
group of aspect-related satellites. These are semantically of two kinds, These satellites require the Ground nominal as prepositional object
indiiting what can be called ‘primary’ and ‘secendary’ aspectual and (in these non-agentive sentences) the Figure nominal as subject.
notions. The primary kind indicate how the action of the verb root is The same holds for the satellite that refers to interiors in the following
distributed with respect to the general flow of time. The secondary kind case:
indicate how the action is distributed with respect to another ongoing
event, namely one of moving along. In translation, these forms can be (89) a. Water poured into the tub
represented as in Table 2.14. - ‘to a point/some points of the inside of’

118 LEONARDTALMY Lexicalization patterns 119
However, EngIish has no form comparable to al1 uver for interiors: b. *Ya v-lil stakan vodoy
(89) b. *Water poured all into/? the tub 1 in-poured glass(Acc) water(rNsrR)
*? poured the glass in with water’
- ‘to all points of the inside of
A new Iocution must be resorted to. This locution, moreover, differs By contrast, the notion ‘al1 round’ (Le. ‘to al1 points of the surrounding
from the others in that it has the reverse valence requirements: the surface of) requires the reversed G-F precedence:
Figure as prepositiona1 object and the Ground (in non-agentive sent-
(95) a. *Ya ob-lil vodu na/? sabaku
ences) as subject:
1 circum-poured water(Acc) on dog(acc)
(90) The tub poured furr uf water *‘I poured water al1 round the dog’
By the opposite token, the satellite for surfaces does not allow this b. Ya ob-fl sabaku v o d o y
reverse valence arrangement: 1 clrcum-poured dog(Acc) water(rNsTR)
‘1 poured the dog round with water’
(91) *The table poured alI over with/of water
This same pattem applies as well to agentive sentences, except that Accordingly, these satellites can be represented notationally as:
what was the subject nominal is now the direct object: (96) a. F. . . +v-(V+ACC>G) b. o.. .fob-(b+msTR>r)
(92) ‘surfaces’ ‘interiors’
a. I poured water onto d. I poured water into the tub Outside Indo-European, Atsugewi exhibits similar cases of Path
the table satellites requiring either basic FG or reversed G-F precedence. Two
b. 1 poured water al1 over e. *1 poured water al1 into such satellites, respectively, are f-c& ‘into a fire’ and +-mik+ ‘into
the table the tub someone’s face’ (represented below as @re and @ce):
c. *1 poured the table al1 f. 1 poured the tub ful1 (97) a . /ach $-s-f-i:-a s-‘-w-ra-pLSa c ah&-P/
over with/of water of water water om-ToPrcALrzERrNFL-pour-afire NP fire-to
3 [Sch se. s&laphiír?.a c %hWP]
Using the earlier notation, the valence requirements of these satelhtes
‘1 poured afire Water-ACC (F) Campfire t0 (G)'
can be represented thus:
‘1 threw water over the campfire’
(93) EI. F . . . *Il (-tO> G) C. F,. . . fin (-tO> G) t-s-‘-i:*”
b- / ach-a? s-‘-w-ra-pl-mik--a
b. F... +d-OVer (fd> G) d. G . , . +full (-Of> F) water-with NONOBJ-TOPICALIZER wrr-pour-aface
c a&tih/
With the concept of a precedence hierarchy among grammatical rela-
Nf man
tions that places subject and direct object above prepositional object,
Js [?ach-á? chea &laphiim-ik- a c %&te]
we can say that in English the notion of a ‘filled surface’ expressed in a
‘I poured aface man-acc (G) water with (F)'
satellite requires the basic Figure-above-Ground, or FG, precedence,
‘I threw water into the man’s face’ (‘1 threw the man
while the notion of a ‘filIed interior’ requires the reverse Ground-above-
aface with water’)
Figure, or GF, precedence.
In many languages, certain notions expressed in satellites requise one
or the other of these san-re precedences. For exampIe, in Russian, the In some cases, a Path satellite can be used with either valence
notion ‘into’ can only be in the basic F-G precedence: precedence. English through works this way in usages like:

a. Ya v-lil vodu v stakan (98) (it= ‘my sword’)
(94)
I in-poured water(Acc) in glass(acc) a. I (A) ran it (F) through him (G)
‘1 poured water into the glass’ b. I (A) ran him (G) through with it(F)

120 LEONARD TALMY Lexicalization paiterns 121

In other cases, there are two satellites, with the same meaning and European languages, sateIlites expressing a ‘ful1 interior’ seem without
sometimes with simiiar forms, that act as a complementary pair in exception to require ,reversed G-F precedence, and satellites expressing
handliig either valence precedence. The Yiddish forms for ‘into’, arayn bounded Paths largely tend to require the Ground as direct object. Are
and ay~, work this way (cf. Talmy 1982): these and comparable pattems language-particular, family-wide, or
universal?
(99) a. Ix hob arayn-geStoxn a dom (F) in ferd (G)
1 have in(StUCk a t h o m in the horse
3.0 Conclusion
‘1 stuck a thom into the horse’
The principal result of this chapter has been the demonstration that
b. Ix hob ayn-gegtow dos ferd (G) mit a dom (F) semantic elements and surface elements relate to each other in specific
I have in(stuck the horse with a thom patiems, both typological and universal. The particular contributions of
‘1 stuck the horse (in) with a thorn’ our approach have included the following:
Certain Russian Path satellites are involved in a further interesting First, the chapter has demonstrated the existence and nature of
valence distinction. They require the Ground as direcr objecf when the certain semantic categories (‘Motion event’, ‘Figure’, ‘Grouud’, ‘Path’,
Path is bounded and is completed ‘in’ a quantity of time. For the ‘precursor’, ‘personation’, etc.) as well as syntactic categories (‘verb
corresponding unbounded Path that lasts ‘for’ a quantity of time, there complex’, ‘satellite’, and ‘satellite-preposition’).
is no Path satellite at al1 but rather a Path preposition that takes the Seccmd, most previous typological and universal work has treated
Ground as prepositional object: languages’ lexical elements as atomic givens, without involving the
semantic components that comprise them. Accurdingly, such studies
(IocJ) a. (i) Satelit obletel zeml’u (v 3 Easa) have been limited to treating the properties that such whole forms can
satellite(rYoM) circum-flew earth(Acc) in 3 hours manifest, in particular, word order, grammatical relations, and case
‘The satellite flew around the earth in 3 hours’ - Le., roles. On the other hand, most work on semantic decumposition has not
made one complete circuit invoived crosslinguistic comparison. The present study has united both
(ii) Satelit letel vokrug zemli (3 dn’a) concems. It has determined certain semantic components that comprise
satellite(nou) flew-along around earth(om) for 3 days morphemes sud assessed the crosslinguistic differences and commonali-
‘The satellite flew around the earth for 3 days’ ties that these exbibit in their pattems of surface occurrence. Thus,
b. (i) On probeial (vs’u) ulicu (v 30 minut) instead of words’ order and role, this study has determined semantic
he length ran al1 street(Acc) in 30 minutes components’ surface presente, site (their ‘host’ constituent or gramma-
‘He ran the Iength of the (whole) Street in 30 minutes’ tical relation), and combination within a site.
(ti) On bezal po ulice (20 minut) Third, our tracing of surface occurrence pattems has extended
he ran along along street(DAT) for zo minutes beyond treating a single semantic component at a time, to treating a
‘He ran along the Street for 20 minutes’ concurrent set of components (as with those comprising a motion event
c. (i) On perebe&l ulicu (v 5 sekund) and its circumstance) . Thus, the issue for us has not just taken the form:
he cross r& street(Acc) in 5 seconds semantic component ‘a’ shows up in surface constituent ‘x’ in language
‘He ran across the Street in 5 seconds’ ‘1’ and in constituent ‘y’ in Ianguage ‘2’. Rather, it has also taken the
(U) On beial Eerez ulicu (2 sekundy) i form: with semantic component ‘a’ showing up in constituent ‘x’ in
he ran along across street(Acc) for 2 seconds and language ‘I’, the syntagmatically related components ‘b’ and ‘c’ show up
potom ostanovils’a there in constituents ‘y’ and ‘z’, whereas Ianguage ‘2' exhibits a different
then stopped surface arrangement of the same ful1 component set. That is, this study
‘He ran across the Street for 2 seconds and then stopped’ has been concemed with whole-system properties of semantic-surface
relations.
The question of universality must be asked with regard .to satellite The present method of componential crosslinguistic comparison
valence distinctions like those we have seen. For example, in Indo- permits observations not otherwise feasible. The following section, 3.1,


demonstrates this for the issue of information’s ‘sdience’. Former Manner or Cause of an event and up to three componen& of a Path
studies of saliente have been limited to considering only whole iexical complex, as in (102).
items and, hence, only their relative order and syntactic roles - and,
appropriate to these alone, have arrived at such notions as topic, (102) The man ran back down into the cellar
comment, focus, old and new information for comparison across In this rather ordinary sentence, English has both packed in and
languages. The present method can, in addition, compare the fore- backgrounded the information that the man? trip to the cellar was
grounding or backgrounding of incorporated semantic components accompIished at a run (ran), that he had already been in the cellar once
accorclmg to the type of surface site in which they show up. It can then recently so that this was a retum trip (back), that his trip began at a
compare the systemic consequences of each language’s selection of such point higher than the cellar so that he had to descend (down), and that
incorporations. the cellar formed an enclosure that his trip originated outside of (in-).
Following this, the Appendix tabularizes and sketches the semantic- Spanish, by contrast, with its different verb-conflation pattern and
surface relations described earlíer, and augments these with a number of &ost no productive satellites, can background only ene of the four
additional categories, to provide a one-glance sense of the relationships English componen& using its main verb for the purpose; any other
that have been uncovered as well as to fumish an expanded ground for expressed component is forced into the foreground in a gerundive or
further research. prepositional phrase. The present example, actually, goes beyond the
issue of how much can be expressed in the background, to that of how
much can be expressed at all in a single sentence, even in the
3.1 The backgrounditzg of meaning in the ver& compiex foreground, without being unacceptably awkward. Here, it tums out,
A theoretical perspective that encompasses both sections I and 2 Spanish can comfortably express either the Manner alone, as in (rqa),
pertains to saliente: the degree to which a component of meaning, due or one of the Path notions together with a gerundively expressed
to its type of linguistic representation, emerges into the foreground of Manner, as in (Igb, c, and d). For acceptable style, any further
attention or, on the contrary, forms part of the semantic background components must either be omitted and left for possible inference, or
where it attracts little direct attention. In this regard, there appears to established eisewhere in the discourse:
be a universal principie. Other things being equal (such as a consti-
(103) Spanish sentences closest to information-packed English sen-
tuent’s degree of stress or its position in the sentence), a semantic
tence of (102)
element is backgrounded by expression in the main verb root or in any
closed-class element (induding a satellite - hence, anywhere in the a. El hombre corrió a-l sótano
verb complex). Elsewhere it is foregrounded. For example, the two the man ran to-the cellar
sentences in (ror) are virtually equivalent in the total inforrnation that ‘The man ran to the cellar’
they convey, but they deber in that the fact of transit by air is pivotd in b. El hombre volvió a-l sótano corriendo
(IoIa) in its nominal (-adverbial) occurrence, whereas it is an incidental the man went back to-the cellar running
piece of background information in (Io1b) where it is conflated within a ‘The man returned to the cellar at a run’
verb. c. El hombre bajó a-l sótano corriendo
(101) a. Last year 1 went to Hawaii by plane the man went down to-the cellar running
‘The man descended to the cellar at a run’
b. Last year 1 flew to Hawaii
d. El hombre entró a-l sótano corriendo
Languages can be quite comparable in the informational content that
the man went in to-the cellar running
they convey. However, a way that languages genuinely differ is in the
‘The man entered the cellar at a run’
amount and the types of information that can be expressed in a
backgrounded way. English and Spanish can be contrasted in this Notice that although the contrast just described was at the leve1 of a
regard. English, with its particular verb-contlation pattee and its general pattem difference between two languages, the same contrast
multiple satellite capability, can convey in a backgrounded fashion the can be observed at the leve1 of individual morphemes, as behveen such

LEONARDTALMY
Lexicaliza tion patterns =25
‘24
If, on the other hand, the English sentence is to achieve equivalence to
Similarly pattemed languages as RUW Iii 1’ ~t~trl English. For example,
the Atsugewi form in backgroundedness of information, then it must
Russian has a Path satehite + prepoglbI&rII ~‘~~lllP’ex~ fp”* k + D*T >
, “rr”‘~d as an mtended destina- drop the ful1 NPS or change them to pronouns, as in:
‘int0 arriYal at’, that characterhes the
Ilmit resort to the Spanish (rg) It blew in
tion. English la& this and, to rende’ 110
,,~r.~rling verb (mive}, with a
pattem of expression using a Path-incbl ” 1 Such equivalence in backgroundlng, however, is only gained by the
Shiku attendant awkwardness in ex1 ,I ,,&“JJ further components of forfeiture of information, for the original Atsugewi form additionally
, , ‘,l”) I’or contra9 are sentences
meting, as seen in (‘o4b). (Shown in ( indicates that the ‘it’ is an icky one and the entry is a Iiquid one.46
exhibiting the usual RussiaeEngfish 1’t’1 “llcliMm)’

a. Russian: On pod-beial k YI” (‘11”” Appendk Compendium of mea&Horm associations
(w)
he up to-ran to p’ ,ll<( IW) This chapter’s research into meaning-form associations is only a begin-
English: He ran up to the bVtd ning. Among other endeavors, it calls for a thorough cross-linguistic
k vt’rotam determination of which semantic categories are represented with what
b. Russ~~: On pri-b&al
he tito ar&&rdlt 1“ fl’itC(DAT) frequencies by which surface constituents. The fine-toothed cataloguing
Engllsh: He arrived at the g”td “t ” run thus called for is initiated here in a more modest format in Table 2.15
W’c can extend the contrast and its annotations. Incorporated there are the semantic-surface occur-
At the general pattem leve1 agai”a rente patterns presented in the text, But these are augmented so as to
,lt,ll cypes of infqrmation they
between languages as to the quantity Mude a number of additional semantic categoríes and one additional
background, for as Engfi& is to Spa”l*h* “o Ats.ugew’ Is t” Engbsh’ verb-complex constituent beyond the verb root and sateilite, namely,
While ti Amerindian Ianguage calll llke Enghsh, backgroundedly verbal inflections.
it can do so as well (as we
indícate Cause and Path ln its verb co’~‘(‘l~~* While the table’s indications are based only on the author’s linguistic
have seen) for Figure and Ground. Tah* , 1 cxample the polysynthetic
i’r
experience and must be amplified by a thorough cross-language survey,
, ,“,ncnted with its morphemes
form in (rgb), here approximateiy t-r 1 such a survey might nevertheless lead to quite few majar upsets:For if a
glossed and separated by dashes: Ianguage comes to attention with a semantic-surface association former-
ly thought non-existent, that association will likeiy be rare. If the table’s
(105) (it) - from-wind-blowing -ick? , ,,lllttwmoved - discrete plus/minus indications are then simply converted to frequency
[Cause.. . . . . . . . .] [Fiylll“ I - ‘l indications, these will exhibit roughly the same pattern as before.
into-liquid - Factual Given such a pattem, the major issue to be addressed next, of course,
Path + Ground is whether the pattem shows any regularities and, if so, what factors
1~ this form in either of two mightexpiain them. The data at hand here suggests only partial
We can try to match English sentenN regularities and, in fact, there are exceptions to every explanatory factor
ways . TO achieve informational equivd INcu* the EngIish sentence must considered. (See Bybee (‘980, 1985) for work on related issues.)
“’ express the additional two
include full independent no-un phra@ However, answers may emerge in the future as more pieces come into
gl¢se NI% can be either accurate
mmponents that it cannot background, place:
,,k in (Ida), or, to equal the
mdicators of the Atsugewi referents, rovide more specific indica- -with the inspection of more languages
ofi@nd f0i-m in colloquialness, they r’H’1 P
tions aat would be pednent t. a pd~~~Cd~~ referent situation, as in -with a more principled determination of which surface forms
(106b). Either way, the mere use of ~‘c’l’ N”s draws atten’on t” the’r are to be considered satellites and how these are to be
contents: distinguished from (say) inflections
-with the inclusion of the remaining verb-complex constituents
(W a. Some icky matter blew into’ sone liquid
#‘ such as adverbial particles and auxiliaries (some of Table
2.15’s semantic categories that are not represented in the
b. The guts biew into the crcfik

126 LEONARD TALMY Lexicalization pattems 127

root, satellite, or inffections, e.g. ‘hedging’ and ‘spatial Table 2.15 (conrd)
Iocation’, are in fact represented in other verb-complex 28. person 1+1 +
- relations to the speaker’s cognitive
constituents) state (namely, to the speaker’s -)
-and with the consideration of further semantic categories and 29. valence/voice (- attention) +
the remaining sentence constituents. 30. factivityjevidence (- knowledge) &)/+ +
31. attitude (- attitude) + -t
32. mood (- intent) +
- relations to the speaker-hearer interaction
Table z-15. Which semantic categories are expressed by which verb- 33. speech-act type (+) +
complex elements G. qualities of the speech event
34. status of the interlocutors [+I + +
semantic categories expressed within the verb-complex by: I-I. factors pertaining neither to the referent
event noc to the speech event
(a) verb root (b) sateflite (c) id ections 35. ‘speaker’s state of mind,
lyesterday’s weather, . :
A. main event
r. main actionlstate + L+/-1 SymboLs used in TabIe 2.15
B. subordinate event + This semantic category shows up in this surface constituent either in many
2. Cause +W) + [+I languages or with great elaboration in at least a few languages.
3 . Marmer +W) (+) This category shows up in this constituent in only a few Ianguages, and there with
f t+, IittIe elaboration.
:: fi%r t-/+1 This category does not show up in this constituent in any languages known to the
C. components of a Motion event author, and may we!l never do so.
6. Figure +W) + l-1 +/- Thlhis ategory shows up in this constituent in one capacity or by one interpretation,
7. Path (and Direction, no. 25) +W) + but not another, as explained in the annotations which follow.
+ [] There is some question about this assignment of f or -, as explained in the
8. Ground alone (+) r-1
7 + 8. Path + Ground +W) + X
annotations which follow.
This category has only slight representation in the verb-compIex constituents
D. essential qutities of the event treated here.
(and of its participants) * Tbis category is possibIy never expressed in the verb-complex constituents treated
g.l hedging 1-l here.
KO. Xdegree of realizatiofl c-1 (+) (M) This category can alone join with the ‘Motion’ category in the verh root and there
II. polarity + form an elaborated system for the expression of Motion events. (IIe category may
12. p h a s e + + $3 also be able to show up in the verb root in other capacities.)
13. =pect
14. Xrate CI CI BRef descriptions and illustrations of SCSMII~~C ategorie~
15. causativity +
16. personation + F+I
17, number in an actor + (+l (Here, (a), (b), (c) refer to the categories’ occurrence in verb roots,
18. distribution of an actor + 11 satellites, or inflections, respectively.)
19. *symmetry/*wlor . . . of an actor -
E. incidental qualities of the event or I. Muin actionlstate. (a) This semantic category - which includes
its participants motion and locatedness - is the one most identified with the verb root. It
20. l relation to comparable events is joined there by the other categories @ven a ‘+’ in column (a). Thus,
21. Xtemporal setting Ll pi in kill, agent causativity (no. 15) joins the main action of ‘dying’ and, in
22. *spatial setting
23. %tatus of the actors pi lie, a Manner notion (no. 3), ‘with a horizontal supported posture’, joins
z+. gender/class of an actor + f the main state of ‘being located’. (b) But there may be an exception to
F. relations of the referent event or its the preceding. By the interpretation favored here for the resultative
pardcipants to the speech event or its participants
25. Direction (deictic)
construction in Indo-European and Chinese languages, the satellite
+U@
26. ‘spatial location (deictic) f-1 fj r-1 presents its expression of a resulting event as the main action or state,
27. tense - +‘ while the verb root, generally expressing a cause, presents this as a

128 LEONARD TALMY Lexicaiization patterns 129

subordinate event. Thus, we consider English melt/rust/rot away to be 5. Result. A causing event (no. 2) always has a resulting event paired
best interpreted as meaning ‘disappear [= +away] by melting / rusting / with it because the two are conceived in terms of a single larger causal
rotting’ and German er-kämpfen/-streiken as meaning ‘obtain [ = +er-] interaction. (a) When both events are expressed together in a verb root,
(e.g. tetitory, wages) by battlinglstriking’. The altemative interpreta- as they can be, the question here is, which event is the main one and
tion would consider the Result expressed by the satellite as the which subordinate? Thus, in I kicked the hall aiong the path, does kick
subordinate event and the verb’s Cause as the main one, with the mean ‘move by booting’ with the Result as main event and Cause as
reading of (say) rusf away then taken to be ‘rust with the result of subordinate, or instead ‘boot with the result of moving’, with the reverse
disappearing’. (c) This category is not indicated by inflections. ascriptions? We favor the former interpretation (the same as in no. 2)-
Thus, it may be that Result never incorporates in a verb root as a
2. Cause. This category refers to the qualitatively different kinds of subordinate event (hence the ‘- ’ in the table in the (a) c&mn), but
causing events such as can be expressed by an English subordinatefiom- only as a main one. (b) In the resultative construction, Result is
or by-clause. It is distinguished from causativity (no. r g), which expressed in the satellite, in many languages with numerous distinc-
corresponds to a superordinate clause of the type ‘NP CAUsEs s’. (a) tions. However, by the interpretation favored here and already discus-
Cause is regularly incorporated in the verb roots of European languages sed in no. r(b), it appears there not as a subordinate event but as the
expressing either motion or other action. Thus, Engliqh blow in The main one. Our conclusion is that all incorporation of Result, whether in
napkin blew ofl the table means ‘move from (due to) the air blowing on verb root or satellite, is as main event. (c) Result is not expressed
[it]‘. (b) Atsugewi has some two dozen prefixal satellites expressing inflectionally.
cause, for example ca- ‘from the wind blowing on [it]‘. (c) Causing-event
types are generally not expressed in inflections. However, by one 6. Figure. The Figure is the salient moving or stationary object in a
analysis, the distinct agentive and inducive inflections of some languages motion event. (a) It is systematically incorporated in Atsugewi’s motion
(e.g. Japanese) do indicate different causing events of the types: ‘[the verb roots, for example in -t’- ‘for a smallish planar object (shingle,
button, stamp, etc.) to move/be-located’. The occasional English
Agent CAUSES s] by acting physically’ VS. ‘. . . by inducing another
examples include rain (It rained in through the window) ‘for rain(drops)
Agent (to act physically)‘.
to fall’. (b) A set of Atsugewi preties, overlapping with the causal set,
3. Manner. Manner refers to a subsidiary action or state that a indicates Figures. A set of Caddo prefures indicates Patient, which
Patient manifests concurrently with its main action or state. (a) It is sometimes coincides with a Motion event’s Figure. (c) Inílections do not
regularly incorporated in most Indo-European languages’ verbs of represent the Figure qua Figure, but they can indicate properties of
Motion (as well as other kinds of action), as in English float in The subject and object - grammatical roles in which the Figure often occurs.
balloon floated into rhe church, which means Lmove, floating in the 7. Parh. This category refers to the variety of paths followed, or sites
process’. (b) Nez Perce has over two dozen prefixal satellites indicating occupied, by the Figure object. (a) It is a regular component in the
Manner, for example %yé*- ‘floating in the process’. (c) Manner is not Motion-verb systems of many language families, for example Polyne-
indicated inflectionally. sian, Semitic, and Romance, as in forms like Spanish entrar ‘move in’,
4. Purpose. A purpose event is one that an agent intends to have salir ‘move out’, subir ‘move up’, bajar ‘move down’, pasar ‘move
occur in consequence of bis undertaking a main event. (a) Purpose past/through’. (b) Path is the main category expressed by the satellites
seems universally excluded from incorporation in Motion verb systems. of most Indo-European languages outside of Romance, as in English
‘Rus, there is no I stored rhe keg into the pantry, with store meaning with forms like in, out, up, down, past, through. (c) Path is not indicated
‘move in order to store’. Pm-pose is incorporated in other type verbs, for inflectionally .
example in wash ‘apply liquid to, in order to clean’ and in hunl (1 hunted 8, Ground. The Ground is the referente-object in a Motion event,
deer) ‘search for, etc., in order to capture’. (b) Purpose is expressed in with respect to which the Figure’s path/site is reckoned. (a) It does not
‘benefactive’ satellites (for example the Atsugewi suf& -iray), which appear alone with the move/be-located component in any language’s
have the meaning ‘in arder to benefitlgive [it] to [the actor named by major Motion verb-root system, but only in occasional forms, like
the direct object nominal]‘. (c) Pm-pose is not expressed intlectionally. English (de-/em-)plane, or in combination with additional components

LEONARD TALMY
Lexicalization patterns 131
130
(see following section 7 + 8). (b) A set of Atsugewi prefures, overlap- falter and leefer as in He teelered on rhe cliff edge, which suggests ‘almost
ping with that for Causes, indicates various body-part Grounds - for falling’. (b) Atsugewi has a suffixal satellite -iwt which indicates ‘almost’
example ‘finger’ or ‘buttocks’ as used with a verb root meaning ‘get a in al1 the customary senses. It is the only such form known to the author.
splinter’. A set of Caddo prefixes indicates Patient, which often (c) This category is apparently not indicated inflectionally.
coincides with a Motion event’s Ground. (c) Inflections do not represent
a Ground object per se but only in so far as it serves as a grammatical II. Polar@. Polarity is the positive or negative status of an event’s
subject or object. existence. (a) Verb roots can incorporate polarity of two kinds, either
that pertaining to the root’s own referent action/state - for example
7 + 8. Path + Ground. The combination of Path and Ground is English hir VS, mLsi; (= not hit) the target - or that pertaining to a
privileged in that it occurs more than other Motion-component com- complement clause’s action/state. In the latter type, incorporated
binations (except for those with the ‘move’ component itself) and polarity even has some of the same syntactic consequences as indepen-
certainly more than the Ground alone. (a) Many languages have a series dent polarity elements (like nor), for example in requiring either some
of verb roots in which this combination joins with ‘move’, for example or any:
English berrh (The ship berthed) ‘move into a berth’ or causative box (1 1 managed to/ordered him to/suspect 1’11 - see someone/*anyone
boxed rhe apples) ‘cause-to-move into a box’. (b) Atsugewi has a major 1 failed to/forbade him to/daubt 1’11 - see anyone/*someone
system of suñixal satellites that express some two-score instances of this (b) Cheyenne indicates the negative with a prefix sáa- in its poly-affrxal
combination, for example -Sr ‘into a liquid’. English has a few verb (Dan Alford, personal communication). (c) Some Amerindian and
examples, such as al@ ‘into the air’, apart (They moved apart) ‘away Asian languages are reported to incorporate positive and negative in
from each other’, and home (1 drove home) ‘to one’s home’. (c) two distinct sets of inflections that otherwise indicate tense, mood,
Itiections do not represent this combination. person, etc.
g- Hedging. Among other functions, hedges qualify the categoriality 12. Phase. Distinguished from aspect because of its different be-
of a linguistic element’s referent. They are mostly indicated around havior, the category of ‘phase’ refers to changes in the status of an
verbs by adverbs or special expressions, like those in He sort of event’s existence. The member notions are ‘starting’ and ‘stopping’, for
danced/He danced ufter a fashion. (a, b,c) However common they may use with any type of event, and ‘initiating’ and ‘finishing’, for events that
be in that form, they are not incorporated in verb roots nor expressed by are intrinsically bounded. To exemplii the two types, I stopped reading
satellites or inflections - unless one considers as hedgers such diminuti- the book refers to a change from reading to not reading at any point in
vizing satellites as Atsugewi -inkiy, which changes ‘to rain’ to ‘to drizzle’ the book, while Ifinished reading the book refers to reading al1 of the
or Yiddish unter-, which in unrer-ganvenen changes ‘to steal’ to ‘to pilfer book, and only then not reading. (a) Phase notions can be incorporated
a bit every now and then’. in verb roots or collocations, as in strike up ‘initiate the playing of [a
IO. Degree uf realization. This category divides a referent action or tune]’ - and, by one interpretation, also in reach (e-g. reach fhe border)
state - almost anywhere along its semantic continuum - into a more ‘finish going toward’, shur up ‘stop talking’, and halt ‘stop moving’.
central core of essential aspects and a periphery of commonly associated Strikingly, ‘stopping’ is expressed only in verbs, like these or as a
aspects, and indicates that-only one or the other of these is realized. complement-taking verb (stop in stop eating) - not as an auxiliary,
Languages regularly indicate this with adverbs or particles near the sateIlite, or inflection. (b) Phase notions other than ‘stopping’ are
verb, for example English aImust and (&sr) barely. Thus, I almost ate ir expressed by satellites, for example ‘finishing’ by German ferrig-, as in
can suggest lifting an item to the mouth and perhaps even inserting and fertig-bauen/-essen ‘fmish building/eating’ (or, more literally, ‘build/eat
chewing it, but excludes at least the essential aspect of swallowing it. to completion’); ‘initiating’ by Gerrnan un-, as in an-spielen ‘open play
Conversely, 1 jurar barely ate it suggests getting an item down the gullet, (e-g. at cards)’ or an-schneiden ‘make the opening cut in’; and ‘starting’
but without the usually attendant gusto in chewing and tasting. (a) It is in the specific sense of ‘bursting out’ by Yiddish be- (+ zix), as in &4.xn
doubtful that a genuine sense of ‘almost’ or ‘barely’ is ever really .zU: ‘burst out laughing’. (c) Depending on the interpretation, phase
incorporated in a verb root. But perhaps coming close are forms like either is or is not expressed in inflections. Thus, a preterite inflection


seems to indicate stopping or finishing in conjunction with an unbound- (intransitive)‘. (c) In Japanese, separate Mections indicate agent
ed or bounded event, as in She s¡ept/She dressed, but may be better causation, inducive causation, and decausitivization.
interpreted as being basically a tensejaspect indicator, ‘wholly occuning 16. Personarion. Personation refers to the configuration of partici-
before now’, that merely implies cessation. There is also inflectional pants that an action is conceived to be associated with. (a) DiBEerent
indication of ‘entry into a state’ - i.e. ‘becoming’ - but it is not clear languages’ verb roots tend to incorporate different personation types.
whether OI not this should be classed together with ‘starting’. Thus, typical for French, the verb for ‘wmb the hair’,peign, intrinsically
13. Aspect. Aspect is the pattem of distribution through time of an refers to one’s doing the action to another (dyadic). The corresponding
action or state. (a) It is regularly incorporated in verb roots, for example Atsugewi verb, c~-s$, refers to one’s manifesting the action in oneself
in English hif, which can refer to a single impact, as against beat, which (monadic). (b) Satellites can reverse a root’s personation type. The
indicates an iteration. (b) It also appears frequently in satelhtes, as in Atsugewi benefactive suffíx makes the ‘wmb’ very dyadic, and the
the Russian prefixal system for indicating perfective/imperfective dis- French reflective - considered here as a satellite - wnverts its verb to
tinctions. (c) It appears regularly in inflections as well, as in the Spanish monadic. (c) Inflections otherwise invohed with causativity may also
conjugational forrns indicating the preterite and imperfect. serve in switching personation types

14. Rate. Rate refers to whether an action or motion takes place 17. Number in un actor. This is the numerosity of th6 participants -
faster or slower relative to some norm. (a) Though some verb roo& from one to many - behaving as any single argument of an event. Ir is
obviously indicate different rates of speed - for example English trudge, listed under category “D” as an essential aspect of an event because
walk, run or nibble, eur, bolr (one’s food) - languages seem to include such numerosity affects how the event is manifested. (a) Many Ameti-
them haphazardly and in conjunction with further semantic differences, dian languages have distinct roots for an action manifested by different
rather than base a genuine system of lexical distinctions on rate alone. numbers of Patients. Thus, the Southwest Pomo verb roots -LV/- ?da/-p%l
(b) Satellites generally appear not to indicate rate, with some potential mean, respectively, ‘for one/two or threejseveral together . . . to go’. It
exceptions: an Atsugewi suf& -i.skur - which has the same form as an is a possible universal that the Patient is the only semantic role
independent verb ‘to hurry’ and, with a verb root, was in elicitation characterized for number in the verb root. (b) It is not clear whether
always translated as ‘hurry up and V’ - might actually or additionally satellites indicate number. The closest case I know is an Atsugewi dual
there mean ‘V quickly’. Dyirbal (Dixon 1972) has a suffix -nbd/-gdiy clitic, -hiy. (c) Inflections in many languages indicate the number of the
said to mean ‘quickly’ but only as part of a semantic range that also subject nominal and sometimes also of the direct object nominal.
includes ‘repeatedly’, ‘start’ and ‘do a bit more’. We have heard one Interestingly, inflectional indications of number seem always to be
report that Yana may have had affixes with precisely the mean@ linked to a particular sytiuctic role, such as subject or object, while
‘quickly’ and ‘slowly’. (c) Rate is not indicated inflectionally. those in the verb root wrrelate instead with a semonlic role, the Patient.

15. CausaMfy. With the notions in this category, an event is 18. Dtitribution of an actor. This refers to the arrangement of
conceived either as occurring by itself or as resulting from another multiple Patients - whether they fonn an aggregate or a linear distribu-
event, where this latter event is either initiated by an agent or not, and tion in space and/or time (in the latter case wrrelating with aspect). (a)
such an agent is either volitional or not. (a) Causative notions are Different distributions are incorporated systematically in certain South-
regularly incorporated in verb roots. Thus, English die indicates only an west Pomo roots: -phil/-hayom ‘for severa1 together/separately to go’,
event of death itself, while murder indicates that a volitional agent has h.srz/pkoy ‘act on objects as a group/one after another’. (b) The
initiated an action that has caused the event. (b) As an example for Atsugewi su& -ayw indicates ‘ene after another’ for multiple Patients.
satellites, the Yiddish prefix far- can be combined with a comparative Though less freely usable, the English satellite oflcan do the same: read
adjective in a verb fonnation meaning ‘to cause to become . . .‘, as in off/check o# (items on a list), (animaIs) die of. (c) There is some
far-beser-n ‘to improve (transitive)’ (from beser ‘better’). If the reflexive indeterminacy as to whether a type of affix like Atsugewi’s -ayw might
zix can be considered a satellite, then it too is an example, for it changes not be better considered inflectional. Other than this, though, inflections
a causative form into a non-causative: farbesern zix ‘to improve seem not to indicate distribution.

134 LEONARD TALMY Lexicaliza fian parterns 135
19. Symmetry, color of an actor. There are many characteristics of an 24. Gender/class of an actor. This refers to category memberships
event’s participants that are not marked anywhere in the verb complex, based on sex or other characteristics, and associated either with an
even though they seem as reasonable (from an a priori perspective) as event’s actors themselves or with the nouns that refer to them. (a) It
the qualities that are marked. Thus, while an argument’s numerosity appears that no verb roots are Iexicalized specifically for use with nouns
and distribution can be marked, there will be no marking for its color or of a particular grammatical gender or class. Thus, for example, Spanish
whether it has a symmetrical arrangement, even though these very could not have two different verbs for ‘to fall’, one for use with
qualities are important in other cognitive systems, such as visual feminine-noun subjects and the other with masculines. While there do
perception . ! exist verb roots associated with nouns of a particular semanfic gender
! (or various other properties), for example roots referring to pregnancy,
20. Relation to comparable everr&. Many adverbial or particle forms
the association seems less one of systematic categotial distinctions
indicate whether an action or state has occurred alone, or in addition to, !
involving selectional features or the like than a matter of individual
or in place of another one of a comparable category, like the forms in
pragmatic applicability; thus, if a man were in fact to become pregnant,
English He oniy dancedi aiso danced/ even danced/ danced instead.
one could simply proceed to say ‘The man is pregnant’. (b) The
These notions, however, seem never to be expressed as satellites or 1
I
grammatical class of the subject and at times also the direct object noun
inflections, or lncorporated in the verb root. is marked by affixal satellites in Bantu languages. (c) The subject’s
I
1 grammatical gender is indicated in the inflections in al1 Hebrew tenses
21. Temporal setting. This category locates an event within a particu-
lar time period, especially a cyclic one. (a) There may be small systems and in the Russian past tense forras, for example in Pos Zayal/Sabaka
of verb roots differing principally as to temporal locale. Thus, English fo layala ‘The hound barked/The dog barked’.
breakf&, brunch, lunch, sup1din.e could be interpreted as meaning ‘to
25. Direction (deictic). This refers to whether the Figure in a Motion
eat in the moming/late morning/midday/evening’. (b) Yandruwanhdha event is moving toward or away from the speaker. (a) It is found
verbs optionally take the suffixal satellites -thalka ‘in the moming’, incorporated in verb roots, for example Englísh come/go and
I
-nhina ‘by day’, or -yukarra ‘at night’ (Bernard Comrie, public presenta-
bring/take. (b) It is frequently marked by satellites, for example the pair
tion). It is possible that only the day’s cycle is ever thus represented and in Atsugewi, -ik./-im, and in Mandarin, . . , Zái/ . . . qù- (c) It is not
not, say, that of the month or year. (c) Inflections appear not to indicate ! marked inflectionally.
this category.
26. Spatial location (deictic). This category would charactetie the
22. Spatial seting. This category would indicate something about the 1
location of an event’s occurrence with respect to the speaker or hearer
physical setting in which an event takes place, perhaps with contrastive (e.g. near or away from one or the other, in or out of their range of
notions like ‘indoors/out of doors’, ‘in the water/on land/in the air’, I
vision). It is readily indicated by adverbs or partides, such as English
‘next to something/in an open space’. But such notions do not seem to here and there. But it appears not to occur otherwise in the verb
I
be marked in our three verb-complex elements. One possible exception ! complex. As possibIe exceptions: we have heard a report that some
is Klamath’s locative suf6xes, though these seem really more to indicate Northwest Coast Amerindian languages have distinct verb roots mean-
Ground than setting, i.e., to,indicate something more like She bit him in ing ‘to be here’ and ‘to be there’, And the evidential satellites or
rhe nose than She bit him in fhe kifchen. The satellites in English eat inffections for visual versus other-sensory information, in Wintu as well
in/eat o~t (suggested by Martin Schwartz) are perhaps a real, if very as other languages, might be used for inferences about spatial deixis.
limited, exception. I
27. Tense. Like the precedigcategory, but for time instead of space,
23. Status of fhe actors. This refers to either absolute or relative social tense characterizes the temporal location of an event with respect to the
characteristics of animate participants in an event. (a) Japanese verbs of moment of the speaker-hearer interaction. (a) By our interpretation,
giving differ according to the relative social rank of the giver and the tense is not incorporated in verb roots. A possible candidate such as
receiver, and so incorporate status. (b,c) Actors’ status does not seem I
English went is considered not as a conflation of semantic ‘go’ t- ‘past’
to appear in satellites or inflections. I
but as a suppletive form standing in the place of the morphemes go and

136 LEONARD TALMY Lexicalization patrems 137

-ed. The reason is that went can only appear in environments where She waslseemed sad. But many verbs indicate state of knowledge
other verb roots are followed by -ed. If went genuinely incorporated a pertaining to a complement event, as in Jan (i) realized/(iiJ conc¡uded
past sense, one might expect its use as well in expressions like *I am that she’d won: (i) the speaker believes the winning to be factual, (ii)
wentirtg = ‘1 was going’, or *I will went = ‘1 will have gane’. (b, c) Tense the speaker is noncommittal about its actual@. (b) Wintu has a set of
is marked by ties and particles (as well as auxiliaries) in many ‘evidentiaI’ suf&es, probably to be taken as satellites, that indicate
languages. It is not clear that any of these should be taken as satellites; whether the speaker knows for sure or infers an event, as wetl as the
the aflixes among them would normally be taken to be idections. evidente by which he atived at his knowledge or supposition (Harvey
28. Persun. Person refers to the relation between an actor in a
Pitkin and Alice Schlichter, conference presentation). (c) In Atsugewi,
referent event and a participant in the speech event (i.e., the speaker or there are two distinct inflectional sets for the ‘factual’ and the ‘inferen-
hearer). Thus, in English, if an actor is the same individual as the tial’.
speaker, the form I is used; if the same as the hearer, you; and if 31. AMude. The category here is the speaker’s attitude toward the
neither, he/she/it or a full nominal is used. (a) No verb roots appear to referent event. (a) Attitude is incorporated in verb roots. For example,
be specific to a particular person. Distinct foros like English am/is the verbs in They raidedlmarauded the village refer to roughly the sarne
invite the same objection as was raised for went above. Japanese verbs objective event, but maraud additionally indicates the speaker’s atiitude
of giving, sometimes suggested as incorporating person, seem rather to of disapproval toward the event. The negative attitudinal content of
basically indicate relative status, which in turn has certain canonic traipse, as compared (say) with walk, is evident from the leadingness of
associations with personal arrangements. (Note that some HUUIZ roots do this question by a trial attomey: Did you confirm that Miss Burnett was
incorporate person, for example the distinct Kikuyu nouns for ‘my tiuipsing around the restaura& (b) The Atsugewi sulXxal sateIlite -inkiy
father’, ‘your father’ and ‘his father’.) (b) If clitics like Spanish me/a indicates the speaker’s ‘cute’ regard for the event. For example, with a
can be construed as satellites, then this part of speech can be given a root ‘flap’, it could be used to speak of baby ducklings moving their
plus for person. (c) Person is notably indicated by inflections. wings about. (c) Attitude seems not to be indicated inflectionally.
zg. Vulence/voice. This category refers to the particular distribution
of attention and perspective point that the speaker assigns to the 32. Mood. Mood refers to a speaker’s feelings or intentions with
aerent actors in an event, when this factor correlates with the surface respect to the actualization of an event. It incIudes a neutral regard, a
cases (grammatical relations) of the nouns referring to the actors. The wish for (something unrealizable), a hope for (something realizable), a
two traditional terms for this category differ only in that ‘voice’ refers to desire to (realize something), and an attempt at (realizing something).
the assignment when it is marked by inflections or auxiliaries and (a) It appears that no verb roots have an intrinsic mood to them. It
‘valence’ otherwise. (a) The category is often incorporated in verb might at first be thought that a verb like wanf, as in She wants fo go, is
roo& for example English sell and buy, which p1ace the main perspec-, desiderative, but it really only refers to the actor’3 desire, not to that of
tive point at the giver and the receiver, respectively, for the same event. the speaker, whose mood toward this event is here neutral. (b,c) Many
(b) The Geman satellite v r- redirects the main perspective onto the languages have a%xes - whether taken as satellites or inffections - that
giver in an exchange, as in ller-kaufen ‘sell’ (VS. kaufen ‘buy’). (c) The indicate mood under terms like indicative, subjunctive, optative, desid-
category is frequently marked by inflections, as in Latin emere ‘to buy’ erative, conative.
and emi ‘to be bought’ . 33. Speech-act qpe. This category indicates the speaker’s intentions
30. Factiv@/evidence. This category distinguishes the speaker’s be- with respect to the hearer in referring to an event. (a) The vast majority
lief in, versus ignorance of, an event’s truth. The two traditional terms, of verb roots are neutral with respect to speech-event type- But a few
factivity and evidente, differ only as to whether this category is verbs do incorporate a particular type, for example the Halkomelem
indicated in the verb root itself or outside it. (a) Only rarely, it seems, roots meaning ‘to be where’ and ‘to go whither’ are solely interrogative,
does a verb root indicate a speaker’s state of knowledge as to its own and mainly imperative are English beware, the collocation be advised
referent event. One example might be English be, indicating speaker’s (which does accommodate modals, but only with an imperative sense:
certainty of a copular attribution, and seem, indicating uncertainty, as in Yuu shouZd/*can be advised rhat . . .), and perhaps forms like whoa,

138 LEONARD TALMY Lexicalizarion patterns 139

giddiyap, scar, (b,c) The category is often marked by satellites and Slobin, Johanna Nichols, Joan Bybee, and Eric Pederson for fruitful
idections. For exampIe, Atsugewi has distinct inflectional paradigms discussions.
2 A zero form in a language ean represent a meaning not expressed by any
for these speech-act types: declarative (1 tell you that . . .), interroga-
tive (1 ask you w h e t h e r . . .), imperative (1 direct you to . , .>, actual lexical item. For example, no German verb has the general ‘go’
meaning of the zero form cited &hen implies walking, so that one could not
admonitive (1 caution you Iest . . .).
ask Wu wollen Sie denn hingehen? of a swimmer).
34. Status of the interlocutors. Status is the same here as in no. 23 but 3 For a further theoretical discussion, exploring questions like: What deeper
refers to the participants of the speech event rather than to the actors of properties of Ianguage can account for why the pattems are as they are?, or
a referent event. (a) The Japanese verbs of giving do not realIy fit here; What still broader phenomena do the observed pattems lit into? see Talmy
(rg?6b, rg?8c, 1983, in preparation).
they basically indicate the actors’ status, and it is only incidental if some
4 Apart from these three processes, an analyst can sometimes invoke what we
of the actors tum out to be participants in a speech event. However, might term semanti resegmenturion. Censider the case of shave as used in
some of Samoa& distinct status-leve1 verbs (e.g. those for eating) may
(0
welI have usages sensitive solely to who it is that is speaking and being
(a) 1 eut John (c) 1 cut myself (e) *1 cut
addressed. (b) Satellites and clitics are used by a number of languages to (b) I shaved John (d) 1 shaved myself (f) 1 shaved
indicate the absolute or relative gender (men’s and women’s speech)
and status of the interlocutors. (c) Inflections for second person in many We could believe here that a reflexive meaning eomponent is lexicalized in
European languages distinguish degrees of formality that are partly the verb, deleted from the sentence, or to be inferred by pragmatics.
based on relative status. However, we need to assume that a reffexive meaning is present only if we
consider this usage to be derived from that iu (b)/(d). We could, altemative-
35. Speaker’s state of mind, . . . It seems that n o markers o r ly, eonclude that the (f) usage is itself basic and refen directly to a particular
incorporations indicate notions unrelated to either the referent event or action pattem involving a single person, with no retlexive meaning at all.
the speech event. If they existed, one might encounter cases like The 5 These forms express universal semantic elements and should not be identied
chair broke-ka meaning ‘The chair broke and I’m currently bored’ or with the English surface verbs used to represent them (they are written in
‘The chair broke and it was raining yesterday’. capitals as MOVE and BEN in other works by the author to underscore this
distinction) .
6 These notions of Figure and Ground have several advantages over Fillmore’s
(e.g. 1977) system of cases. The comparison is set forth in detail in Talmy
NOTES
(r978a), but some major differences can be indicated here. The notion of
I Grateful ackuowledgement is here extended to several people for their ‘Ground’ captures what is common - namely, a funetion as referente-objeet -
native-speaker help with Ianguages eited in this chapter: to Selina LaMarr for to al1 of Fillmore’s separate cases ‘Location’, ‘Source’, ‘Goal’, and ‘Path’,
Atsugewi (the language of the author’s fieldwork studies), to Mauricio Mixto wbich otherwise have nothing to indicate their commonality (as against, say,
and Carmen Silva for Spanish, to Matt Shibatani and to Yoshio and Naomi ‘Instrument’, ‘Patient’, and ‘Agent’). Further, Fillmore’s system has nothing
Miyake for Japanese, to Tedi Kompanetz for French, to Vicky Shu and to indieate the commonality of its Source, Goal, and Path cases as against
Teresa Chen for Mandarin, to Luise Hathaway and Ariel Bloch for German, Location, a distintion captured in our system by the move/be-loeated
to Esther Talmy and Simon Karlinsky for Russian, and to Ted Suppala for opposition within the Motion component. Moreover, the Fillmorian cases’
Ameriean Sign Language. incorporation of path notions (together with referente-object fnnction)
In addition, thanks go to several people for data from their work on other opens the door to adding a new case for every newly recognized path notion,
languages: to Wallace Chafe for Caddo, to Hamo Aoki for Nez Perce, to with possibly adverse ccnsequences for universality claims. Our system, by
Robert Oswalt for Southwest Pomo, to Horma Gerdts for Halkomelem, to abstraeting away all notions of path into a separate ‘Path’ component, allows
Ariel Bloch for Arabic, to Bradd Shore for Samoan, and to Elissa Newport for the representation of semantic complexes with both universal and
and Ursula Bellugi-Klima for Ameritan Sign Language - as well as to several language-particular portions.
others whose personal communications are eited in the text. The author has 7 The assessment of whether i? is Manner or Cause that is conflated in the verb
supplied the Yiddish forms, while the Latin data are from dictionaries. is based on whether the verb’s basic referente is to what the Figure does or to
Special thanks go to Tim Shopen for his invaluable editorial work with earlier what the Agent or Instrument does. For exampIe, in ‘1 rolled the keg . . .‘,
drafts for this chapter. And thanks as well to Melissa Bowerman, Dan roiled basically refers to what the keg did and so expresses Manner, whereas


in ‘1 pushed the keg . . .‘, refers to what 1 did, and so gives the Cause expresses (in one area of its usage) the curling together of the fingers of a
of the event. hand specifically caused by intemal (neuromotor) activity. No other cause
8 There appear to be constraints, some of them apparently universal, on the can be compatibly expressed in conjunction with this verb:
kinds of subordinate-ciause material that can systematicahy conflate with a. My hand clenched into a fist from a muscle spasm/*from the wind
(3
Motion. For example, while English readily cordlates the Manner or the blowing on it
Cause into a main Motion event, as demonstrated in (6), it cannot do so for
b. I/*He clenched my hand into a t%t
the Purpose. Thus, beside the complex sentence with subordinate Pm-pose
clause in (ia), there exists no conflated analog like (ib): By contrast, CWI ~p expresses a main action similar to that of clench, but it
a . I moved the keg into the pantry, in order to sto;e it incorporates no restrictions as to the cause of the action:
(9
b. *I stored the keg into the pantry (vi) a. My hand curled up into a 6st from a muscle spasm/from the wind
blowing on it
In English, then, a non-main event is allowed to conflate with a mam Motion
event if its time of occurrence is before or during that Motion event, but not b. I/He curled my hand up into a fist
after it. This restriction may well be universal. g The systematic relations of the kind shown here are discussed with greater
Under language-particular constraint, on tbe other hand, is another type of detall and rigor in Tabny (1975). But one point from that discussion can be
non-main event, Lprecursor’, which expresses an action &eady undergone by brought in here. A distinction must be made between tmnslational motion
the Figure object that is currently in Motion. Thus, Atsugewi does regularly and contained motion. In the former, an object’s basic Iocation shifts from
exhibit forms analogous to (iib) with meanings like that in (lia): one point to another iu space. In the latter, the object keeps its same basic, or
(3 a. I moved (put) the blanket into the basket, having first folded it ‘average’, location. This latter case usually involves rotation, oscillation,
expansion/contraction, or ‘contained wander’. This distinction in types of
b. 1 folded the blanket into the basket
motion underlies the analysis shown in (6) for rol1 and bomce. Botb these
In English the sentence in (iib) would have to be interpreted as expressing verbs iu their non-translational sense refer to motion, but only to contained
concurrent Manner: folding the blanket in ?he process uf piacing it in. motion, as seen in (a) and (b).
However, English does minirnally exhibit the precursor pattem with sent-
ences like Reuch/Get me the salt, meaning ‘Give me the Salt, having first (a) The log rolled over and over in the water
reached to/gotten it’, orto the extent that speakers accept sentences like He (b) The hall bounced up and down on the pavement square
scooped up some jeliy beans into my bag in the sense of ‘He put into my bag
These verbs, like float in the sense of ‘be afloat’, can then take on the
some jelly beans that he had scooped upjhaving first scooped them up’.
additional meaning of translational motion tbrough space.
Aside froti any systematic conflation with Motion tbat they may or may
IO Tbe usage relationships posited here are accorded some psychological real@
not exhibit, certain types of non-main events do otherwise show up conl3ated
by data on children’s errors. Bowerman ( 198 I ) documents a stage in English
in main verbs. Purpose, for example, LF conflated in the non-motion-system
acquisition where children become ‘aware’ of motion conflation in verbs and
English verbs wash and rinse. These verbs, beyond referring to certain
then overextend the pattem. Thus, verbs that in adult English, idiosyncrati-
actions involving the use of liquid, indicate that such actions are undertaken
cally, cannot be used with an incorporated motion rbeaning become so used
in arder fu remove dirt or soap. Evidente for such an incorporation is that the
b y children:
verbs are virtually unable to appear in contexts that pragmatically confhct
witb the Pm-pose: I (a) Don’t hug me off my chair (= by hugging move me off)
(iii) a. I washed/rinsed the shirt in tap water (b) When you get to her [a doll], you catch her off (on a merry-go-
b. *1 washed/rinsed the shirt in dirty ink round with a doll, wants a friend standing nearby to remove the
do11 on the next spin around)
whereas otherwise comparable verbs like soak andfIr& which seem not to
express any Purpose beyond the performance of the main action, can appear (c) 1’11 jump that down (about to jump onto a mat floating atop the tub
there: water and force it down to the bottom)

(iv) I soaked the shirt in dirty ink/I fhrshed dirty ink through the shirt Further support comes from historical changes in Word meaning, which
exhibit similar extensions beyond previous category boundaries. ThW in
Further, Cause and Manner can be conflated as wel1 in verbs that do not
participate in the Motion system. For example, the English verb clench their traditional use the ver& hold and carry formed a near p-fea suppletive

LEONARDT A L M Y Lxxicalization patterns 143

pti, differing only in that carry additionaIIy incorporated a Motion event d . .., rNsrRvMRN?ed-to-break . . .= . . . rbroke . . .
while hold did not: e. . . . Auruoxed-to-break . . . = . . . *,broke . . .
without motion with motion f. . . . AGENred-to-break ... = . . . ,broke . . .
1 held the box as 1 lay on the *1 held the box to my neighbor’s g. . . . UNDERWENT-tO-break . . = . . . Dbroke . . .
bed llOUS2
‘1 carried the box as 1 lay on 1 carried the box to my neigh- The causing event can be expressed not onIy by a full clause, as in the text
the bed bor’s h o u s e examples, but also by a verb-derived nominal, as in (b) below, or by what can
be termed an ‘action noun’, as in (c). A standard noun as in (d), however,
Currently, though, carry in some contexts (those where motion just has will not do:
occurred or is about to occur) can also be used in a locative sense: Istood irá
fioti of the door canyiug the box. (While the children’s examples extended The window cracked -
non-motion verbs to motion usages, this case has gone in the opposite (a) from a balI’s sailing into it- - nominabzed clause
direction.) (b) from the pressure/bump of a branch
II This regular conflation of motion with path in these languages seems rarely against it - verbderived nominal
to extend to any regular conflation of location with site - Le., to any basic
system of distinct verb roots expressing ‘be-in’, ‘be-en’, ‘be-under’, etc. Such (c) from the wind/a fire/the rain - action noun
a system has recently been reported (Donna Gerdts, personal communica- (d) *from a hall - standard noun
tion) for Halkomelem, a Salish language of Canada. But generally, these
languages instead use some single form, roughly expressing ‘be-at’, in The clause-like behavior of action nouns can be attributed to their being in
conjunction with a series of adpositions - much like English when no Manner fact conflations of ful1 clauses. Thus, the examples in (c) might be considered
has contlated with be, as in The peri wus (rather than lay) in/on/under rhe to have intemal semantic stmcrures equivalent to the following clauses:
box. wind ‘air’s blowing [on the Figure]’
Though rarely forming a core system, the verbal expression of location + rain ‘rainwater’s falling [on the Figure]’
site is clearly under no prohibitory constraint. English, for one, has a number tire ‘dames acting [on the Figure]’
of incidental instances of such conflation, for example surround (‘be
around’), top (‘be atop’), flank (‘be beside’), adjoin, span, line, fil1 - as in A Such semantic wnflation, taking place in the noun, exemplifies Iexicalization
ditch surrounded the field, A cherry topped the desseti, Clothing filled the in a grammatical category other than those, the verb root and the satellite,
hamper. It is just that such verbs seidom constitute the colloquial system for addressed in this chapter. (For further examples, involving conflation in
locative expression. subordinating and wordinating wnjunctions and in certain adverb classes,
12 This is not to imply that a verb root always has exactiy one basic aspect. A see Talmy rg?8b).
verb root can show a certain range of aspects, each manifesting in a different 15 It is not only intransitive sentences that can be autonomous. For example, An
context. Thus, English kneel is one-way in She knelt when the bell rang and is ucorn hit the piare is autonomous. The requirement, rather, is that the
steady-state in She knelt there for a minute. sentence must not express a cause (as does An acorn broke the plate).
13 These two grammatical forms - ‘keep -ing’ and ‘Vdummy u [- + Derivj,’ - may 16 Arguments are given in Talmy (rg76a, r978b) why the resulting-event
be thought to tigger certain cognitive processes. Respectively, these are (b) form should be considered semantically more basic than the causing-
‘multiplexing’ and ‘uuit-excerpting’. Such processes are discussed in Talmy event (c) form.
(197W 17 This impinging object is the Figure within tbe causing event, but it is the
14 Other linguistic treatments (e.g. McCawley 1968) represent their incorpo- Instrument with respect to the overa11 cause-effect situation. That is, for this
rated causative element by the capitaliied form ‘ CAUSE’. Since more distinc- autbor ‘Instrument’ is not a basic notion as it is, say, for Fillmore (1975). It is
tions are recognized here, more representational forms would be needed a derived notion, to be characterized in terms of other, more basic notions:
(and are in fact used in the author’s other writings): the Instrument of a cause-effect sequence is the Figure of the causing event.
18 The act of will is the 6rst link in the causal chain. Through intemal
a. . . . broke . . . =.. . broke . . . (neuromotor) activity, it brings about the movement of the body. Note that
b. . . . REsuLred-to-break ... = . . . Rbroke . . . such bodily motion, even when not referred to, is a necessary link for a final
c. . . . EvENred-to-break . . . = . . . ubroke . _. physical event. Thus, while Abe burnt the Leaves only mentions Abe as the

Lxxicalizalion patterns 145
144 LEONARD TALMY

initiator and the leaves’ buming as the final event, we must infer not only that 23 Verbs that range over two lexicalization types can be used either with or
frre was the immediate Instrument but also that Abe (due to bis will) acted without a grammatical augment for the same meaning. We see this for hid.s
physically to marshal it. over the agentive and self-agentive types, and for set . . . won over the
19 To describe this more analytically: something acts on a sentient entity, self-agentive and inducive types:
causing within it the intention to carry out an act. The intention in tum leads
to its actually carrying out the act, in the usual manner of agency. Thus, the (a) She hid herself behind the bushes = She hid behind the bushes
entity is caused to act as an Agent (so that another good term for the (b) He had bis dogs set upon (i.e. fall upon) us = He set his dogs upon
‘inducive’ is ‘caused agency ‘) . us
The act that is referred to in most inducive verbs is a self-agentive one, and
in particular one of ‘going’, e.g. smoke (oti) ‘by applying smoke, induce to 24 For these, the three aspect-causative types we have noted for verbs of state
go (out)’ (atypically , sic/set . . . on refer to an agentive act of ‘attacking’). have the following particular manifestation: (a) a body or object is in a
Because most self-agentive verbs are intransitive like most autonomous verbs posture non-causatively , or else an animate being self-agentively maintains
(the other verb types require a direct object), an inducive construction its body in the posture; (b) a body or object comes into a posture
relates to a self-agentive one in much the same way that an agentive non-causatively, or else an animate being self-agentively gets its body into
construction relates to an autonomous one: the posture; (c) an agent puts a body other than its own, or some other
(a) inducive: (b) self-agentive: object, into a posture.
25 The stative usage of the Iast two verbs here may not be immediately obvious.
They sent the drunk out of the bar. The drunk went out of the bar
(c) agentive: (d) autonomous: It can be seen in the following:
They threw the drunk out of tbe The drunk sailed out of the bar
bar (a) She bent over the rare flower for a ful1 minute
(b) He bowed before his queen for a long minute
There seems to be a corresponding kind of semantic ‘drift’: we tend to
understand a self-agentive event as occurring in and of itself, and to take the
26 The pattem we are concemed with here held better in older forms of English.
inducer of an inducive event as directly bringing about the ha1 event without
Thus, the idea of agent derivation for the verb is quite questionable for
the intermediary volition of the actor.
20 It is, however, quite possible that no verbs distinguish between the (c) and modem English. But enough of the pattem remains to serve as illustration
(d) causation types, even cross-linguistically, so that these would have to be and to represent languages that do have such forms clearly. Among these
latter are apparently many Uto-Aztecan languages (Wick Miller, personal
merged. The (a) and (b) types are distinguished perhaps only in the stative,
communication) and Halkomelem.
as in English by the verbs be and stay:
27 This use of the reflexive is a special grammatical devine, not a SemanticaIly
(a) Tl-re pen was on the incline (autonomous situation) motivated one, because there is no way to construe the normal meaning of
(b) The pen *was/stayed on the incline from a lever pressing against it the reflexive in this context. NormalIy, the reflexive entails that exactly what
(resulting-event causation) one would do to another, one does to oneself. In the present case, what one
does to another is to place one’s arms around his/her body, lift, and set
21 We can avoid the problem with tilay - that it is bi-morphemic, with a prefix down. But that is clearly not what one does with oneself. The movement is
explicitly expressing unintentionality - by using the verb spill in a pair with accomplished, rather, by intemal- i.e., neuromuscular - activity.
polrr. Tbis same pair would aiso allow illustration of the ‘S . . . too . . .’ 28 This sufiix in Spanish generally incorporates a passive meaning (unlike the
frame, which misiay/hide do not easily fit: I spilied/*poured the milk by otherwise comparable Japanese -te, which has no voice characteristics).
opening the spout toa wide. However, the present construction, as in estaba acostado - which might be
22 The same test frames employed here can also be used witb verbs hke break, taken literally as ‘1 was laid-down’ - will genera@ be understood with a
that can incorporate any of a range of causative types, to select out one non-passive reading, as in the sentetice gloss ‘I lay (there)‘.
particular causative reading. For example, break is interpretable only as an 2g As noted earlier, it is somewhat forced for modem English to interpret
author type verb in (a) and only as an agent type in (b): posture verbs as pure statives, with augmentation required for the other
aspect-causative types. For one thing, marking of an agentive-non-agentive
(a) 1 broke the window by pressing against it toa hard distinction has al1 but disappeared ColIoquially, with forms like lay or sir
(b) 1 broke the window in arder to let the gas escape serving for both meanings. For another, the satelIite can often appear in


stative usages as well. Thus, the combination of verb + satellite can to a large 33 Note that actions lacking physical contact can also be lexicalized with
degree be used equally for all three aspect-causative types: different personations. For example, the English verb get (‘go and bring
(a) He laid down/stood up all during the show back’) is basicalty monadic but can add a benefactive expression for the
dyadic. On the other hand, the roughly equivalent serve (‘bring to someone’)
(b) He laid down/stood up when the show began is basically dyadic but can add a reflexive for the monadic type (the reflexive
(c) She laid him down/stood him up on the bed here signals only this change in personation type, for it lacks the literal
Nevertheless, a distinction in the use of forrns does still hold to this extent: interpretation it has in I shaved John/I shaved myselfl:
the satellite seems somewhat awkward in some stative expressions, for monadic dyadic
example in He laid (?down) therefor Tours. And the verb without satelhte is I got some dessert from the kitchenj1 got some dessert from the kitchen
somewhat awkward in colloquial speech for the agentive usage: ?She for Sue
iaidjstood the Child on the bed. 1 served myself some dessert from( 1 served Sue some dessert fiom the
30 The postures category is mostly non-relational. One can largely determine a the kitchen kitchen
body’s coniiguration by observing it alone. But the ‘positions’ category is
relational. It involves the position assumed by one object with respect to 34 For this section, the earlier Iimitation to single-morpheme verbs has been
another (especially where the latter provides support). Some position notions relaxed. Considered here, thus, are a lexical complex like rip ofland, later, a
that are frequently found lexicalized in verbs across languages are: ‘he on’, morphemically complex verb like frighten. This is feasible because valence
‘stand on’, ‘lean against’, ‘hang from’, ‘stick into’, ‘stick out of, ‘stick/adhere properties can inhere in morphemic complexes of this sor-t as well as in single
to’, ‘float on (surface)‘, ‘float/be suspended in (medium)‘, ‘be lodged in’, roots.
‘(clothes) be on’, ‘hide/be hidden (from view) + Loc’. The postures and 35 Actually, this paradigm is abridged from a still larger one (see Tahny
positions categories may have no clear boundary behveen them or may 197~301-375) that distinguishes three Figure-Ground precedence relations:
overlap. But these heuristic ciasses, in some version, do seem to be treated the basic format with Figure above Ground in the case hierarchy, that with
differently in many languages. Figure demotion alone, and that with Figure demoted and Ground pro-
31 English does have a few instances where a lexical item, unhke hide, can moted. Perhaps no singIe verb exhibits all the forms, but a pair of verbs can
participate in expressions for all three state relations, including state- serve to illustrate (cp. FiIImore (r977), Hook (1983)):
departure:
non-agentive agentive
(a) She stoud there speaking basic precedence
(b) She srood up to speak The bees swarmed in the 1 pounded my shoe on the table.
(c) She stood down when she had finished speaking garden .
with Figure demoted
32 To be sure, English has m- and de-/dis- for use with some position and
condition verbs (unload, decentralize). But their use is Iimited, and it is also It swarmed with bees in the 1 pounded with my shoe on the
Iargely secondary in that the forms indicate reversal of state-entry rather than garden. table .
state-departure directly. (Thus, central must first add -ize indicating state- and with Ground promoted
entry before it can add de-; there is no *decentral.) The garden swarmed with bees. 1 pounded the table with my shoe.
The distinct treatment that languages accord state-departure as against
state-location and state-entry often shows up as well in their adpositional Note that the with appearing here a s a demotion particle and still marking the
systems expressing Path. For example, the same morpheme expresses ‘at’ Figure becomes the with that marks the Instrument when a sentence of the
and ‘to’ but a Merent one expresses ‘from’ in French à/a/de, Japanese present sort is embedded in a causative matrix (6. note 17). Thus, the
ni/ni/kara (though e is also used for the ‘to’ meaning), and Atsugewi sentence in (a) can be embedded as in (b) to yield (c):
-ip/-i?/-uka. English exhibits this pattern in some of its prepositional and (a) 1 kicked the ball (G) with my left foot (F)
relative-interrogative forms: [cI kicked my left foot (F) into the bah (G)]
(a) She was behind the bam Where was she? (b) 1 *Movad the hall (FJ across the field (~2)
(b) She went behind the bam Where did she go? by kicking it (G1) with my left foot (Fo)
(c) She came from behind the barn Where did she come from? (c) I kicked the hall (F) across the field (G) witb my Ieft foot (Fo j 1)

LEONARD TALMY Lexicalizaiion patterrts 149
148

36 The final genitive expression here would now be only literary. However, forms named next in the text, such as English verb particles, Latin veib
there are other verbs that take a colloquial mit phrase containing the Figure: prefixes, Chinese resultative compléments, and the non-inflectional affíxes in
the Atsugewi polysynthetic verb. Probably also deserving satellite status are
(a) Ich warf faule Apfel auf ihn Ich bewarf ihn fallen Apfeln such compound-forming verbal adjuncts as the first element in English (to)
‘I threw rotten apples at hhn’ ‘1 pelted him with rotten apples’ test-drive; on the’other hand, free adverbs, even ones related semantically to
(b) Ich schenkte ihm das Fahrrad Ich beschenkte ihn mit dem the ,verb root rather than (say) to the whole clause, seem less like satellites.
Fahrrad AIso seeming to merit satelbte status are the incorpoiated nouti in Iroquoian
‘1 “presented” the bicycle to ‘1 “presented” him with the polysynthetic verbs, whereas pronominal clitics as in French seem.less to do
him bicycle’ so and full noti phrases are entirely excluded. What status sbouId be
acwrded such verbphrase forms as a negative element or closed-class
37 Tbe two valence types here pertain not only to verbs but also to adjectival particles like English only and even is uncertain. It is not clear whether this
and larger constructions that express affect. Thus, the expressions italicized indeterminacy is due to the present theory’s early stage of development or to
below can be used only with the case-frame surround shown for them: a cline-like character for the satellite category. :
Stimuhs 42.5 subject Experiencer as subject 4~ There appears to be a universal tendency toward sat&te formation:
That ti odd to me 1 am glad about that elements .with certain types of meaning tend to leave the loc+ions in a
That is of importance to me 1 am in fear of that sentence where tbey logi&ly belong and move into the verb complex. ‘Phis
That got the goat of me-t gol 1 jIew off the handie over that tendency, whose extreme expression is polysynthesis, is also regularly
my goat evident in smaller degrees. Examples in English are the negative and other
38 Enghsh used to favor Stimulus-subject even more than it does now, but a emphatic modifiers on nouns:
number of verbs have shifted their valence type. For example, the affect (a) *Nut Jóan hit himJ JÓan didn’t hit him
ved rue and like - as well as the sensation verb hunger and the cognition (b) Even Joan hit him j J óan even bit him
verb think - used to take the Experiencer as grammatical object but now take
(c) Joan gave him on!y one+Joan oraiy gave hirn óne
it as subject.
3g These lists avoid verbs that refer more to an affect-related action than to the 43 Not al1 Path expressions permit omissions of.this so+ Such is the case with
affect itself. For example, quuke and ranf - candidates for the Experienccr- up to in the sense of.’ ‘approach’ and also with into in the sense of ‘colli&on’:.
subject group - really refer directly to the subject’s overt actions, and oniy (a) When 1 saw Joan on the comer, 1 walked up to her (‘. . . walked
impiy his/her accompanying affect of fear or anger. Similarly, hnrass and up)
placafe - potentially Stimulus-subject verbs - refer more to the activities of (b) It ds too dark to see the tree, so be walked into it (*. . . walked
an externa1 Agent than to the resultant state of ir-ritation or calm in the
b)
Experiencer.
40 Phis arrangement applies as well to verbs of sensation. Thus, ‘be cold’ is
4.4 Judging from their di&bution, satellites of this type seem to be an areal
phenomenon rather than a genetic one. Thus,. Atsugewi and Klamath,
lexicalized from the point of view of the Experiencer feeling the sensation.
-Ah& is added for the perspective of the Stirnulus object rendering the neighboring but unrelated languages, both have extensive s$xal systems of
these, satellites. But the Pomo languages,. related to Atsugewi and sharing
sensation:
(a) verb root: -yi:skap- ‘feel cold’ with it the extensive instrumental prefix system (see section 2.5), quite lack
inflectional &-set: .g ‘_ w- -= ‘I - subject (3d person-object)’ Path + Ground satellites.
/s-‘-w-yi:skap -“/ * [&ye&tph]
45 Though this may remove some of Atsugewi’s mystiqbe; notice that the
‘1 am cold (i.e., feel cold)’ German satellite entgegen- also has the ‘in going to meet’ meaning, as in
entgegenlaufen b-un to meet’, while Latin ob- parallels Atsugewi -ikc still
{b) verb root: -yi:skap- ‘feel cold further in having both the ‘meeting’ and the ‘passage-blocking’ meanings, as
valence-shifting sufl%: -ah& ‘from Stimulus to Experiencer’ in occurrere ‘run to meet’ and obstruere ‘build so as to block &f’:
inflectional affix-set: ‘- w- -a ‘3d person-subject’ 46 The Atsugewi polysynthetic verb can background.still more: Deixis and four
/‘-w-yi:skap-ah&-“/ 3 [&ye&páh&a] additional nominal roles - Agent, Inducer, Companion, and Beneficiary.
‘It is cold (i.e., to the touch)’ However, Deixis is distinguished only as .between ‘hither’ and ‘hence’,, and
41 Tbere is some indeterminacy as to exactly which kinds of constituents found the nominal roles only as to person and number or, in certain circumstances,
in construction with a verb root merit satellite designation. Clearest are the merely their presente in the referent situation.

Talmy lexicalizationpatterns

More Related Content

What's hot (20)

Viewers also liked (9)

Similar to Talmy lexicalizationpatterns (20)

Recently uploaded (20)

Talmy lexicalizationpatterns