![Interrogating corpora to describe
grammatical patterns
Kristin Davidse, *Lieven Vandelanotte, Caroline Gentens
and Lobke Ghesquière
KU Leuven (University of Leuven) and *University of Namur
Corpus linguistics achieved its first major successes by drawing new types of in-
formation about lexical items from large scale corpora. This was the result of in-
novative thinking, implemented into creative corpus interrogation tools. From
the 1970s on Firth’s (1957: 196) insight that the meaning of a word is defined by
its collocates was developed into a corpus methodology (e.g. Jones & Sinclair
1973; Sinclair 1991) that was to revolutionize the field of lexicosemantics. One of
its first landmark achievements was the Collins COBUILD English Language
Dictionary (Sinclair 1987), a dictionary created from scratch on the basis of sys-
tematic study of concordances for each entry. For each lexical item, this allowed
the identification of its major subsenses, their most important collocate ranges,
their semantic prosody, their near-synonyms and antonyms, and the grammati-
cal, or colligational, contexts they occurred in. All these dimensions of lexicose-
mantic corpus research are currently being developed further not only in corpus
linguistics (e.g. Stubbs 1995, 1996), but also in cognitive linguistics (e.g. Barlow &
Kemmer 1994; Tummers, Heylen & Geeraerts 2005) and more formal approaches
(e.g. Fellbaum 2004; Fellbaum et al. 2005), and in cognitive psychology (e.g. De
Deyne & Storms 2008). The collocational view of lexical meaning is, for instance,
informing advanced work on semantic categorization, in which partial synonymy
is established on the basis of similarity between collocate clouds (De Deyne,
Peirsman & Storms 2009).
A similar convergence between theoretical innovation and new corpus
methodologies occurred from the 1990s on, when the idea that lexicon and gram-
mar form a continuum (e.g. Sinclair 1991; Goldberg 1995, 2006) was translated into
specific corpus methodologies (e.g. Hunston & Francis 2000; Stefanowitsch & Gries
2003). What these efforts have in common is that they make “the study of grammar
more similar to the study of the lexicon, [...] and more amenable to investigation by](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-16-320.jpg)
![Interrogating corpora to describe grammatical patterns
understood subject of please in the first sentence, which is a control construction,
but the understood object of please in the second one, which is a tough construc-
tion. He focuses on the case of ready, which untypically allows the formation of
both constructions and can in principle lead to instances ambiguous between the
two, as in the chicken is ready to eat. He points out that the tough construction can
be investigated only if its passive variant which avoids the ambiguity, e.g. the
policeman stood ready to be questioned, is brought into the picture. He then exam-
ines the role of semantic and textual features in the choice between the active and
passive variant, as manifested in historical data slices from Late Modern to Pres-
ent-day English. Rudanko and Rickman re-examine from a corpus-based perspec-
tive another issue framed within formal grammar. They look at the sentential
complements of warn, a verb of the influence type (Sag & Pollard 1991), i.e. one
that describes a participant influencing another participant to perform the action
in the sentential complement, e.g. I would warn her against paying exorbitant pric-
es for books and objects of art. These are control constructions, as the direct object
of the matrix sentence is the controller of PRO in the sentential complement. Ac-
cording to Bach’s Generalization (Rizzi 1986), the object NP must be structurally
represented in object control structures. However, the authors show that warn is
attested in covert object control constructions. In a corpus study that looks at the
last two centuries, they investigate whether the covert NP object receives a general
or a specific interpretation, how it relates to register, and how this developed over
time. Rudanko also revisits the hairy issue of the different meanings of the to in-
finitival and of -ing complement of the adjective afraid. He singles out Wierzbicka’s
(1988) insightful treatment, which associates “tentative intention” with the to in-
finitive, e.g. she was afraid to wake her mistress up, and the dreading of what “might
happen” with the of -ing form, e.g. he was afraid of going insane. He investigates the
comparable semantic contrast between [+choice] for the to infinitive and [–choice]
for the of -ing complement in relation to authentic corpus data. Maier studies the
distribution of pronoun case forms in it-clefts and it BE sentences, environments
in which the occurrence of subject pronouns besides object pronouns has long
puzzled scholars. He proposes the hypothesis that the subject pronouns have been
reanalysed as postverbal focus markers through markedness reversal, and he links
the empirical verification of this hypothesis in corpus data to different variables,
set up as necessary and sufficient conditions. Rohdenburg’s study of dual form
intensifiers such as real – really starts from the hypothesis that their distribution is
parallel to that of simple comparative forms and ones with more, e.g. realer – more
real. This hypothesis is confirmed: the simpler forms favour attributive adjectives
and non-complemented adjectives while the heavier ones favour non-attributive
and complemented adjectives. His datasets allow him to assess whether this distri-
bution is prosodically, formally or semantically motivated.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-20-320.jpg)
![ Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière
References
Barlow, M. & Kemmer, S. 2000. Usage-Based Models of Language. Stanford CA: CSLI.
Biber, D., Johansson, S., Leech, G., Conrad, S. & Finegan, E. 1999. Longman Grammar of Spoken
and Written English. London: Longman.
De Deyne, S. & Storms, G. 2008. Word associations: Network and semantic properties. Behavior
Research Methods 40: 213–231.
De Deyne, S., Peirsman, Y. & Storms, G. 2009. Sources of semantic similarity. In Proceedings of
the 31st Annual Conference of the Cognitive Science Society, N. Taatgen & H. Van Rijn, H.
(eds), 1834–1839. Austin TX: Cognitive Science Society.
Fellbaum, C. 2004. Idiome in einem digitalen lexikalischen System. Linguistik und Literatur 34:
56–71.
Fellbaum, C., Delfs, L., Wolff, S. & Palmer, M. 2005. Word meaning in dictionaries, corpora, and
the speaker’s mind. In Meaningful Texts: The Extraction of Semantic Information from
Monolingual and Multilingual Corpora, G. Barnbrook, P., Danielsson & M. Mahlberg (eds),
31–38. Birmingham: Birmingham University Press.
Firth, J.R. 1957. Papers in Linguistics 1934–1951. London: OUP.
Francis, G., Hunston, S & Manning, E. 1996. Collins COBUILD Grammar Patterns, 1: Verbs.
London: HarperCollins.
Francis, G., Hunston, S & Manning, E. 1998. Collins COBUILD Grammar Patterns, 2: Nouns and
Adjectives. London: HarperCollins.
Fried, M. 2010. Grammaticalization and lexicalization effects in participial morphology: A Con-
struction Grammar approach to language change. In Formal Evidence in Grammaticaliza-
tion Research [Typological Studies in Language 97], A. Van linden, J.C. Verstraete & K.
Davidse (eds), 191–223. Amsterdam: John Benjamins.
Goldberg, A. 1995. Constructions: A Construction Grammar Approach to Argument Structure.
Chicago IL: University of Chicago Press
Goldberg, A. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford:
OUP.
Goldberg, A. & Jackendoff, R. 2004. The English resultative as a family of constructions. Lan-
guage 80: 532–568.
Hunston, S. & Francis, G. 2000. Pattern Grammar: A Corpus-driven Approach to the Lexical
Grammar of English [Studies in Corpus Linguistics 4]. Amsterdam: John Benjamins.
Jackendoff, R. 1975. Tough and the trace theory of movement rules. Linguistic Inquiry 6:
437–447.
Jones, S. & Sinclair, J. 1973. English lexical collocations. Cahiers de Lexicologie 24: 15–61.
Postal, P.M. 1971. Cross-Over Phenomena. New York NY: Holt, Rinehart and Winston.
Postal, P.M. 1974. On Raising. Cambridge MA: The MIT Press.
Rizzi, L. 1986. Null objects in Italian and the theory of pro. Linguistic Inquiry 17: 501–57.
Sag, I. & Pollard, C. 1991. An integrated theory of complement control. Language 67: 63–113.
Sinclair, J. et al. (eds). 1987. Collins COBUILD English Language Dictionary. London: Collins.
Sinclair, J. 1991. Corpus, Concordance and Collocation. Oxford: OUP.
Stefanowitsch, A. & Gries, S. 2003. Collostructions: Investigating the interaction of words and
constructions. International Journal of Corpus Linguistics 8: 209–243.
Stubbs, M. 1995. Collocations and semantic profile: On the cause of the trouble with quantita-
tive studies. Functions of Language 2: 1–33.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-25-320.jpg)
![Interrogating corpora to describe grammatical patterns
Stubbs, M. 1996. Text and Corpus Analysis: Corpus-Assisted Studies of Language and Culture.
Oxford: Blackwell.
Stubbs, M. 2009. Technology and phraseology: With notes on the history of corpus linguistics.
In Exploring the Lexis-Grammar Interface [Studies in Corpus Linguistics 35], U. Römer &
R. Schulze (eds), 15–32. Amsterdam: John Benjamins.
Tummers, J., Heylen, K. & Geeraerts, D. 2005. Usage-based approaches in Cognitive Linguistics:
A technical state of the art. Corpus Linguistics and Linguistic Theory 1: 225–261.
Wierzbicka. A. 1988. The Semantics of Grammar [Studies in Language Companion Series 18].
Amsterdam: John Benjamins.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-26-320.jpg)
![ Patricia Ronan
Against this background of studies carried out on contemporary and earlier
varieties, let us now move on to consider the developments in data which is
sampled on one coherent principle from the Old English to the Early Modern
English period.
3. Data and method
The data for the present study has been obtained by manual and semi-automatic
searches of ca. 50,000 words each of Old English, Middle English, and Early Mod-
ern English texts from mixed genres. For the Old English period, data has been
collected from the Toronto Dictionary Corpus in Electronic form (diPaulo Healy
et al. 2004), and the data is a subset of a ca. 108,800 word data set that contrib-
uted to an earlier study (Ronan 2012a). It represents a genre mix of poetry, histor-
ical-religious narrative in translation from Latin, original narrative and annalistic
evidence (see Table 2). The Middle English data mainly stems from the Helsinki
Corpus (Kytö 1996 [1991]) and has been supplemented by data from the General
Prologue and The Knights Tale from Chaucer’s Canterbury Tales (de Castro, n.d.).
In addition to the latter, original religious narrative has been used in the form of
data from The Book of Margery Kempe and short legal material from the Statutes
of the Realm. Data from the Early Modern Period has been taken from the
Helsinki Corpus only. The intention has been to collect in the corpus texts all
instances of semantically low-content verbs that are predicated by predicate noun
objects, and in which the resulting verbal group forms a single semantic unit.
Thus, the predicate nouns are either action nouns, or abstract nouns whose col-
location with verbs could be paralleled by simple verbs, e.g. make a story or do
evil. The existence of a corresponding simplex at the period of attestation is not
considered a strict criterion, however, as simplexes may have come into existence
only later, such as in the case of a/to battle, or, conversely, they may have existed
in the past as in the case of to take wife, to which Old English wīfian correspond-
ed, but was lost later. Clearly, even if they are not at the time of attestation, light
verb constructions may at some point in time be paralleled by a simplex. If any
doubt existed as to whether a structure should be considered a light verb con-
struction, inclusion has generally been opted for. Overtly aspectual or causal
verbs, such as onginnan/gan ‘begin, start’, have however been excluded in this
study. These sampling considerations have continuously been followed in all pe-
riods under scrutiny.
Particularly in more literary genres from the Middle English period onwards,
a doubling of predicate nouns was repeatedly found. In order not to distort the](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-35-320.jpg)
![ Patricia Ronan
Table 2. Old English corpus texts.
Text Words Genre Period Light verb frequencies/
100,000
Anglo-Saxon
Chron A
15 000 Chronicle 9th-10th century 187
Beowulf 12 000 Poetry 9th – 10th century 333
Ohthere and
Wulfstan
2500 Travel narrative 10th century 120
Bede I (extract) 16 500 Historical narrative 10th-11th century 133
Old English total 45000 207
(7) ...swa deorlice dæd gefremede fagum sweordum
‘He performed a valuable deed with shining swords.’ (Beo [0162 (583)])
(8) .. ac ymb Hreosnabeorh eatolne inwitscear oft gefremedon.
‘But often committed dreadful malicious slaughter at Hreosnabeorh.’
(Beo [0681 (2472)])
Fremman seems to correspond to later dōn and macian collocations and denotes
active implementation of the action described by the predicate noun. Fremman is
used in 15 different types, and particularly help and dæd ‘deed’ show notable token
frequencies at 4 and 3 respectively. The predicate nouns may either have an equiva-
lent verb or not, though the latter is rarer in the sample. There are no examples of
fremman in the samples from Chronicle A or Ohthere and Wulfstan. This may in-
dicate that fremman is used more in poetic diction than in non-literary contexts.
Semantically, fremman shares some space with wyrcan ‘cause, work’, which is also
used in the corpus material. While wyrcan might express causation more than im-
plementation in some examples, in others the parallels to fremman are strong, e.g.
(9) þær is mid Estum an mægð þæt hi magon cyle gewyrcan
‘The Estonians have the ability that they can create cold’
(Or1 [0148 (1.17.33)])
The parameters that govern this interaction need further research on the basis of a
larger corpus. From the sample texts it appears that in addition to semantic con-
tent and individual stylistic preferences, alliteration between light verb and predi-
cate nouns also favours the use of light verbs constructions.
Next in frequency in the data sample is the light verb niman ‘to take’. This too is
found with some instances in Bede I, but shows highest token frequency in the
Chronicle A, where it particularly appears in apparently set expressions like sige ni-
man ‘take victory’ and freod nimman ‘make peace’ (10 examples each). In contrast
to fremman, the type frequency for niman in the sample is low, but the tokens appear](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-37-320.jpg)
![Light verb constructions in the history of English
repeatedly, which will be at least partly due to the subject matter of the Chronicles,
which are of course much concerned with the topics of victory and peace.
(10) þy ilcan geare gefeaht Ecgbryht cyning Beornwulf cyning on Ellendune,
Ecgbryht sige nam
‘That same year king Ecgbryht and king Beornwulf fought at Ellendune
and Ecgbryht took victory.’ (ChronA [029810 (823.2)])
In spite of its high attestation numbers in these cases, niman may not be entirely
grammaticalized in the contexts, as alternative structures exist with habban, as in
(11) below. By and large, the collocations with niman indicate or emphasize that
the agent benefits from the predicate noun.
Habban ‘have’ is a further high-frequency light verb in the sample corpus.
Akimoto Brinton (1999: 50) point out that one of the effects of using this light
verb with predicate nouns is that using an otherwise obligatory object of an action
noun may be avoided, as in cyððe habban ‘to have knowledge’ as opposed to X
cyðan ‘make X known’. They further stress that in the nominal construction, the
predicate noun may be modified by an adjective, and that this is not possible with
the simple verb. The collocation of habban and predicate noun typically expresses
a state or a result state as illustrated by
(11) hi sona wið heora feondum gefuhtan, sige hæfdan
‘they soon fought with their enemies and had victory.’ (B1 0134 (9.44.28)].
This example illustrates the resultant state of peace. The example is also interesting
because sige ‘victory’ has no corresponding simple verb. That this is not of paramount
importance for the use of the light verb construction, however, is shown by the fact
that its synonym sib, which is paralleled by the simplex sibbian ‘to make peace’, is also
used in light verb constructions with habban. A survey of the instances of light verb
constructions in the Old English sample corpus is given in Table 3 below.
Overall, in the sample texts, the frequency of the light verb constructions was
highest in the poetic text of Beowulf with 40 tokens corresponding to a frequency
of 333 per 100,000 words. This is followed by the historical Chronicle A, whose 28
tokens correspond to a frequency of 187 per 100,000 words. There is a clear drop
in the narrative genres with a frequency of 133 in Bede I (22 tokens) and Ohthere
and Wulfstan at a frequency of 120, albeit on the basis of only few tokens (3 tokens).
This suggests that genre of the text, stylistically elevated or aiming for exactness of
expression, is a more important incentive for the use of the construction in Old
English than is the question of whether the text is translated from Latin or not. An
overview of the attestations of light verb constructions in the corpus are provided
in Table 3.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-38-320.jpg)
![ Patricia Ronan
Table 3. Frequencies of light verb constructions in the Old English sample.
Verb Types Tokens Frequency per 100,000
Fremman 15 23 51
Niman 4 22 49
Habban 15 18 40
Wyrcan 8 9 20
Don 7 7 16
Sellan 7 7 16
Limpan 2 3 7
Beran 1 2 4
Dælan 1 1 2
Giefan 1 1 2
Total counts 61 93 207
4.2 Middle English data
The Middle English data set for this study consists of approximately 51,000 words
taken from different genres. The source texts from the Helsinki Corpus are The
Book of Margery Kempe, and the roughly contemporary legal materials from the
Statutes of the Realm II from the Helsinki Corpus (Kytö, 1996 [1991]). Two ex-
amples of Chaucer’s late 14th century Canterbury Tales, The Prologue and the
Knight’s Tale are taken from the eChaucer corpus (Ne Castro, 2007). Texts and data
are tabulated in Table 4.
In the Middle English sample texts, the line-up of light verbs has changed
considerably from the Old English sample. The most frequently used light verbs
are have (114), make (61), do (45), take (26), and give (18). Further examples are
found with get, go, grant, hold, bring, perform, say, work, arrange, bear, commit,
exercise, find, hold, pursue, purvey and put.
Compared to the Old English data, both the raw number of tokens, as well as
the number of different types of predicate nouns used with the light verbs of high-
er frequency has increased. Particularly make, take and give have gained ground,
Table 4. Middle English corpus texts.
Text Words Genre Period Light verb frequencies/
100,000
Chaucer 8422 + 21872 Poetry –1380–1400 494
Kempe 9670 Narrative 1436 672
Statutes II 11 271 Legal 1420–1500 790
MidE total 51235 595](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-39-320.jpg)
![ Patricia Ronan
(16) By which Co[m]myssions, and auctoritie yeven to the seid Co[m]myssioners
in the seid fourme, meny greate hurtes and inconvenyences in dyv[er]s p[ar]
ties of this yor realme doon and had by encresse of Water were necessarily
redressed refourmed and amended...
‘Many great hurts and inconveniences done and had in diverse parts of
your realm by increase of water were necessarily redressed, reformed and
amended by the commissioners and the authority given to the said com-
missioners in the said form.’ (Statutes II 68–72)
(17) so sche dede hir massage as sche was comawndyd.
‘And so she did her errand as she was commanded.’ (Kempe 218–19)
While in some cases stylistic considerations like rhyme, rhythm or other prag-
matic considerations may cause the use of the light verb, in this last example no
clear alternative involving a simple equivalent to do a message seems to exist.
The new light verbs in this period, make and take ostensibly replace fremman
and niman, and give replaces sellan, which had become more restricted in meaning
after the OE period. Though it has a more recent history as a light verb than do,
make has overtaken it in the sample texts from the Middle English period, where
it is already well represented with 38 tokens, though this might look different in a
larger corpus. Statutes II particularly use it in the context make feoffment ‘grant a
fief’, for which the simplex feoffen also existed (OED, sv feoff). Otherwise, make
mention is well attested, and make act and make list are also found repeatedly.
(18) þe creatur of whom þis tretys makyth mencyon
‘The creature of whom this treatise makes mention.’ (Kempe 456–7)
There are also further examples with rarer verbs which continue to be medium
frequent or rare support verbs in contemporary English, such as grant or put (in)
(Allerton, 2002, pp. 186, 190), and an example of commit is in evidence:
(19) ... where the Attourney named in the seid Warrant put the seid Article in
execucion
‘where the attorney named in the said warrant put the said article in
execution’ (Statues II 628–9)
(20) ... all man~ of accions [...] to graunt to ev~y of such p~sones p~teccion
[...]
‘all manner of actions to grant protection to all such persons’
(Statutes II, 292–4)
(21) Forasmuche as Thomas Crofte commytted a detestable murdre within the
Marches of Wales
‘Forasmuch as Thomas Crofte committed a detestable murder within the
Marches of Wales’ (Statutes II, 829)](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-41-320.jpg)
![Light verb constructions in the history of English
Present Day English light verbs are already well represented in the Middle English
corpus data, but in Old English there is only a partial overlap with Present Day
English verbs, which means that a large percentage of Old English structures is
overlooked in that approach. The present study has also asked the question of why
the numbers of attestation may be the highest in Middle English corpus data and
it is suggested that language contact may have played an important role in the
increase in frequency. However, this assumption, as well as the observed frequen-
cies, should be tested on considerably larger corpora comprising more varied cor-
pus texts and genres.
References
Akimoto, M. Brinton, L. 1999. The origin of the composite predicate in Old English. In Col-
locational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton
Akimoto (eds), 21–58.
Algeo, J. 1995. Having a look at the expanded predicate. In The Verb in Contemporary English.
Theory and Description, B. Aarts C.F. Meyer (eds), 203–217. Cambridge: CUP.
Allerton, D. 2002. Stretched Verb Constructions in English. London: Routledge.
Brinton, L. 1996. Attitudes to increasing segmentalization. Complex and phrasal verbs in Eng-
lish. Journal of English Linguistics 24: 186–205.
Brinton, L. M. Akimoto. (eds). 1999. Collocational and Idiomatic Aspects of Composite Predi-
cates in the History of English. [Studies in Language Companion Series 47]. Amsterdam:
John Benjamins.
Claridge, C. 2000. Multi-word Verbs in Early Modern English. Amsterdam: Rodopi.
Danchev, A. 1992. The evidence for analytic and synthetic developments in English. In History
of Englishes: New Methods and Interpretations in Historical Linguistics, M. Rissanen, O. Ih-
alainen, T. Nevalainen I. Taavitsainen (eds), 25–41. Berlin: Mouton de Gruyter.
Denison, D. 1981. Aspects of the History of English Group-Verbs: With particular Attention to
the Syntax of the Ormulum. PhD dissertation, University of Oxford.
diPaulo Healy, A., D. Haines, J. Holland, D. McDougall, I. McDougall X. Xiang (eds). 2004.
The Dictionary of Old English Corpus in Electronic Form. TEI-P3 conformant and TEI-P4
conformant version. Toronto: DOE Project.
eChaucer: NeCastro, G. 2007. eChaucer.http://www.umm.maine.edu/faculty/necastro/chau-
cer/ (29 May 2012)
Goldberg, A. Jackendoff, R. 2004. The English resultative as a family of constructions. Lan-
guage 80: 532–568.
Hiltunen, R. 1983. The Decline of the Prefixes and the Beginnings of the English Phrasal Verb [An-
nales Universitatis Turkuensis Series B, Vol. 160]. Turku: University of Turku Press.
Hiltunen, R. 1999. Verbal phrases in Early Modern English. In Collocational and Idiomatic As-
pects of Composite Predicates in the History of English, Brinton Akimoto (eds), 133–166.
Hock, H.H. Joseph, B. 1996. Language History, Language Change, and Language Relationship:
An Introduction to Historical and Comparative Linguistics. Berlin: Mouton de Gruyter.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-48-320.jpg)
![ Patricia Ronan
Hoffmann, A. 1972. Die verbo-nominal Konstruktion – eine spezifische Form der nominalen
Ausdrucksweise im modernen Englisch. Zeitschrift für Anglistik und Amerikanistik 20:
158–183.
Kytö, M. (ed.) [1991]1996. Manual to the Diachronic Part of the Helsinki Corpus of English Texts.
Coding Conventions and Lists of Source Texts, 3rd edn. Helsinki: Department of English,
University of Helsinki.
Kytö, M. 1999. Collocational and idiomatic aspects of verbs in Early Modern English. In Collo-
cational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton
Akimoto (eds), 167–206.
Matsumoto, M. 1999. Composite predicates in Middle English. In Collocational and Idiomatic
Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 59–96.
Matsumoto, M. 2008. From Simple Verbs to Periphrastic Expressions. Frankfurt: Peter Lang.
Nickel, G.1968. Complex verbal structures in English. IRAL - International Review of Applied
Linguistics in Language Teaching 6: 1-21.
Nurmi, A. 1999. A Social History of Periphrastic DO. Helsinki: Société Néophilologique.
Orchard, A. 2003. A Critical Companion to Bēowulf. Cambridge: D.S. Brewer.
Poussa, P. 1990. A contact-universals origin for periphrastic do, with special consideration of
OE-Celtic contact. In Papers from the 5th Internatioanl Conference on English Historical
Linguistics [Current Issues in Linguistic Theory 65], S. Adamson, V. A. Law, N. Vincent
S. Wright (eds), 407–434. Amsterdam: John Benjamins.
Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the
English Language. London: Longman.
Rissanen, M. 1991. Spoken language and the history of do-periphrasis. In Historical English
Syntax, D. Kastovsky (ed.), 321–342. Berlin: Mouton de Gruyter.
Ronan, P. 2012a. Make peace and take victory. Support Verb Constructions in Old English in
Comparison with Old Irish. NOWELE Supplement Series 24. Odense: University of South-
ern Denmark Press.
Ronan, P. 2012b. Mobilizing linguistic concepts: support verb structures in Early English. In
English on the move: Mobilities in literature and language [SPELL 27], D. Britain A. Kern-
Staehler (eds), 183–99. Tübingen: Narr.
Tanabe, H. 1999. Composite predicates and phrasal verbs in the Paston Letters. In Collocational
and Idiomatic Aspects of Composite Predicates in the History of English, Brinton Akimoto
(eds), 97–132.
Tieken Boon van Ostade, I. 1990. The origin and development of periphrastic auxiliary do: A case
of destigmatisation. North-Western European Language Evolution (NOWELE): 16: 3–52.
Traugott, E.C. 1999. A historical overview of complex predicate types. In Collocational and Idi-
omatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds),
239–260.
Wierzbicka, A. 1982. Why can you have a drink when you can’t *have an eat? Language 58:
753–99.](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-49-320.jpg)
![ Stefan Diemer
seems quite a fanciful phrase, but it is not used figuratively: in the underlying
manuscripts, distance becomes indeed a variable, with particles and articles at-
taching to and detaching from their respective partners. Examples (4) to (8) illus-
trate this variability. The illustrations are from Wycliffite texts written around 1380
AD, thus from the (for prefixation late) period Middle English 3 (ME3, as per the
Helsinki Corpus designation). The by far largest share of prefixed verbs follow es-
tablished prefix patterns, as shown in example (4) with the verb forsake, the same
usage as in Present-day English.
(4) forsake
A man schal forsake fader [...]
‘A man shall forsake father [...]’ (MS Bodley 296 1v, detail)
Example (5) shows the prefixed verb totake, in which the prefix to is written in a
slightly different font and at about half the usual distance between words.
(5) totake
totake marie þi wyf [...]
‘to take Mary your wife [...]’ (MS Junius 29 13v, detail)
Example (6) shows the non-lexical prefix a- separated from -shamed, creating a
rather unusual group of two non-lexical morphemes forming one discontinuous
lexeme.
(6) a schamed
and þei weren not a schamed [...]
‘and they were not ashamed [...]’ (MS Bodley 296 1v, detail)
Example (7) illustrates that even articles could be prefixed, raising the question of
whether a morphosyntactic analysis of article prefixation in Middle English could
produce qualitative differences between different types of articles.
(7) þeday
“þei were before to þeday [...]”
‘they were before to the day [...]’ (MS Douce 370 5v detail)](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-53-320.jpg)
![What happened to the English prefix?
Example (8) shows the preposition into written as two separate particles.
(8) clepe in to (PDE: call sb.)
god: clepe in to adam [...]
‘God called Adam [...]’ (MS Douce 370 6r detail)
These manuscripts also illustrate the advantage of a hand-written text: if a scribe
was not sure whether to separate the particle or not, he could muddle the waters
and semi-separate it, as example (9) shows with two different versions of into.
(9) in to
(MS Laud misc 361, lv, and MS Bodley 665, 13v, details)
Diachronic corpora (such as the Helsinki Corpus), for the most part, do not dis-
tinguish these forms; if the data is tagged, these stages of removal are not uni-
formly categorized. In view of these taxonomic problems it is not surprising that
most research on the morphosyntactic category of prefixes has been qualitative or
based on small quantitative samples, which may explain the wide range of answers
to the question why the prefix disappeared. It is not easy to discuss the disappear-
ance of the prefix on a global scale, as various prefixes may behave differently.
Furthermore, a quantitative analysis is difficult for the reasons mentioned above.
3. The decline of the prefix: Theories
Broadly speaking, three different hypotheses are proposed to explain the decline of
the prefix:
1. lexical change and competition
2. syntactical change
3. semantic change](https://image.slidesharecdn.com/3385057-250601225200-375692d3/85/Corpus-Interrogation-And-Grammatical-Patterns-Kristin-Davidse-54-320.jpg)
Corpus Interrogation And Grammatical Patterns Kristin Davidse
- 1. Corpus Interrogation And Grammatical Patterns Kristin Davidse download https://ebookbell.com/product/corpus-interrogation-and- grammatical-patterns-kristin-davidse-6770114 Explore and download more ebooks at ebookbell.com
- 2. Here are some recommended products that we believe you will be interested in. You can click the link to download. Corpus Linguistics And Second Language Acquisition Perspectives Issues And Findings Xiaofei Lu https://ebookbell.com/product/corpus-linguistics-and-second-language- acquisition-perspectives-issues-and-findings-xiaofei-lu-44993206 Corpus Inscriptionum Iudaeaepalaestinae No 704 Hannah Cotton https://ebookbell.com/product/corpus-inscriptionum-iudaeaepalaestinae- no-704-hannah-cotton-47205106 Corpus Inscriptionum Iudaeaepalaestinae Vol2 Hannah Cotton https://ebookbell.com/product/corpus-inscriptionum-iudaeaepalaestinae- vol2-hannah-cotton-47205108 Corpus Inscriptionum Iudaeaepalaestinae 3 Walter Ameling Hannah M Cotton https://ebookbell.com/product/corpus-inscriptionum- iudaeaepalaestinae-3-walter-ameling-hannah-m-cotton-47205110
- 3. Corpus Inscriptionum Iudaeaepalaestinae Hannah Cotton https://ebookbell.com/product/corpus-inscriptionum-iudaeaepalaestinae- hannah-cotton-47205116 Corpusassisted Ecolinguistics 1st Edition Robert Poole https://ebookbell.com/product/corpusassisted-ecolinguistics-1st- edition-robert-poole-47208538 Corpus Linguistics For Writing Development Philip Durrant https://ebookbell.com/product/corpus-linguistics-for-writing- development-philip-durrant-47229108 Corpus Exploration Of Lexis And Discourse In Translation Meng Ji https://ebookbell.com/product/corpus-exploration-of-lexis-and- discourse-in-translation-meng-ji-47642756 Corpusassisted Discourse Studies Mathew Gillings Gerlinde Mautner https://ebookbell.com/product/corpusassisted-discourse-studies-mathew- gillings-gerlinde-mautner-48269930
- 6. Corpus Interrogation and Grammatical Patterns
- 7. Volume 63 Corpus Interrogation and Grammatical Patterns Edited by Kristin Davidse, Caroline Gentens, Lobke Ghesquière and Lieven Vandelanotte SCL focuses on the use of corpora throughout language study, the development of a quantitative approach to linguistics, the design and use of new tools for processing language texts, and the theoretical implications of a data-rich discipline. For an overview of all books published in this series, please see http/benjamins.com/catalog/scl Studies in Corpus Linguistics (SCL) General Editor Elena Tognini-Bonelli The Tuscan Word Centre/ The University of Siena Consulting Editor Wolfgang Teubert University of Birmingham Advisory Board Michael Barlow University of Auckland Douglas Biber Northern Arizona University Marina Bondi University of Modena and Reggio Emilia Christopher S. Butler University of Wales, Swansea Sylviane Granger University of Louvain M.A.K. Halliday University of Sydney Yang Huizhong Jiao Tong University, Shanghai Susan Hunston University of Birmingham Graeme Kennedy Victoria University of Wellington Michaela Mahlberg University of Nottingham Anna Mauranen University of Helsinki Ute Römer Georgia State University Jan Svartvik University of Lund John M. Swales University of Michigan Martin Warren The Hong Kong Polytechnic University
- 8. Corpus Interrogation and Grammatical Patterns Edited by Kristin Davidse Caroline Gentens Lobke Ghesquière University of Leuven Lieven Vandelanotte University of Namur/University of Leuven In collaboration with Tinne van Rompaey University of Leuven John Benjamins Publishing Company Amsterdam/Philadelphia
- 9. Library of Congress Cataloging-in-Publication Data International Conference on English Language Research on Computerized Corpora (33rd : 2012 : Leuven) Corpus Interrogation and Grammatical Patterns / Edited by Kristin Davidse, Caroline Gentens, Lobke Ghesquière and Lieven Vandelanotte. p. cm. (Studies in Corpus Linguistics, issn 1388-0373 ; v. 63) This volume is a selection of strictly refereed and extensively revised papers from the ICAME 33 International Conference “Corpora at the Centre and Crossroads of English Linguistics” organised in Leuven from 30 May to 3 June 2012. Includes bibliographical references and index. 1. English language--Grammar--Data processing--Congresses. 2. English language- -Research--Data processing--Congresses. 3. English language--Discourse analysis--Data processing--Congresses. 4. Computational linguistics-- Congresses. I. Davidse, Kristin, editor. PE1074.5.I5 2014 420.1’88--dc23 2014021646 isbn 978 90 272 0371 7 (Hb ; alk. paper) isbn 978 90 272 6974 4 (Eb) © 2014 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa 8 TM Cover design: Françoise Berserik Cover illustration from original painting Random Order by Lorenzo Pezzatini, Florence, 1996. The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.
- 10. Table of contents Acknowledgements vii List of contributors ix Interrogating corpora to describe grammatical patterns 1 Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière part 1. Patterns in the verb phrase Light verb constructions in the history of English 15 Patricia Ronan What happened to the English prefix, and could it stage a comeback? 35 Stefan Diemer The pattern to be a-hunting from Middle to Late Modern English: Towards extrapolating from Wright’s English Dialect Dictionary 57 Manfred Markus The present perfect and the preterite in Late Modern and Contemporary English: A longitudinal look 81 Johan Elsness can and be able to in nineteenth-century Irish English: A case of ‘imperfect learning’? 105 Marije van Hattum part 2. Patterns in the noun phrase Syntactic constraints on the use of dual form intensifiers in Modern English 131 Günter Rohdenburg
- 11. Corpus Interrogation and Grammatical Patterns Ma daddy wis dead chuffed: On the dialectal distribution of the intensifier dead in Contemporary English 151 Zeltia Blanco-Suárez The case of focus – The reanalysis of subject pronouns as focus markers in subject predicative complement position 173 Georg Maier part 3. Patterns in complementation structures Null objects and sentential complements, with evidence from the Corpus of Historical American English 209 Juhani Rudanko and Paul Rickman A new angle on infinitival and of -ing complements of afraid, with evidence from the TIME Corpus 223 Juhani Rudanko Active and passive infinitive, ambiguity and non-canonical subject with ready 239 Mikko Höglund part 4. Patterns of clause combining The diffusion of English absolutes: A diachronic register study 265 Nikki van de Pol and Hubert Cuyckens It-clefts in English L1 and L2 academic writing: The case of Norwegian learners 295 Hilde Hasselgård The speech functions of tag questions and their properties. A comparison of their distribution in COLT and LLC 321 Ditte Kimps, Kristin Davidse and Bert Cornillie Auther index 351 Subject index 355
- 12. Acknowledgements This volume is a selection of strictly refereed and extensively revised papers from the ICAME 33 International Conference “Corpora at the Centre and Crossroads of English Linguistics”, organised in Leuven from 30 May to 3 June 2012. The con- ference was organized by Kristin Davidse (University of Leuven) and Lieven Van- delanotte (University of Namur) with the invaluable help of Tinne Van Rompaey (University of Leuven). When Tinne was not able to participate in the editing of the proceedings, Kristin and Lieven were very pleased that Caroline Gentens and Lobke Ghesquière joined the editorial team. Wout Van Praet helped with the for- matting and proofreading, and with the compiling of the indexes. While naturally we accept the final responsibility for the choices made, we were very lucky in being guided by the conscientious advice of the following exter- nal referees: David Banks, Marcus Callies, Claudia Claridge, Hendrik De Smet, Hans-Jürgen Diller, Marion Elenbaas, Jonathan Fine, Jennifer Herriman, Risto Hiltunen, Rika Ito, Thomas Kohnen, Bettelou Los, Ronald Macaulay, Christian Mair, Ilka Mindt, Lilo Moesner, Britta Mondorf, Terttu Nevalainen, Gerard O’Grady, Javier Pérez-Guerra, Peter Petré, Julia Schluter, Monika Schulz, Peter Siemund, Erik Smitterberg, Sali Tagliamonte, Olga Timofeeva, Jukka Tyrkkö, Uwe Vosberg, Susanne Wagner, Regina Weinert, Debra Ziegeler. We much appreciated their careful comments and suggestions for revision, which benefited both the papers selected for this volume and those eventually recommended to other pub- lication outlets. Without the contributors, there obviously would not have been a volume. We thank them for developing the studies they presented at ICAME 33 with an eye to the general coherence of this volume and for the good spirit in which the whole enterprise was conducted. John Benjamins helped along the project with their usual friendly efficiency. Kees Vaes was always ready to offer help and useful advice and we are very grateful to Elena Tognini-Bonelli for having accepted the volume for publication in the Studies in Corpus Linguistics Series. For generous financial support of the conference, thanks are due to the Re- search Foundation – Flanders (FWO), the Fonds de la Recherche Scientifique (FNRS) and the Leuven International Doctoral School for the Humanities and
- 13. Corpus Interrogation and Grammatical Patterns Social Sciences. We would also like to acknowledge the support of the GOA- project 12/007, The multiple functional load of grammatical signs, awarded by the Leuven Research Council and coordinated by Kristin Davidse, Bert Cornillie, An Van lin- den and Jean-Christophe Verstraete. Kristin Davidse, Caroline Gentens, Lobke Ghesquière, Lieven Vandelanotte
- 14. List of contributors Zeltia Blanco-Suárez Universidade de Santiago de Compostela Department of English and German, Facultade de Filoloxía Rúa de San Roque, 2 E-15782 Santiago de Compostela, Spain Hubert Cuyckens University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Kristin Davidse University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Stefan Diemer Saarland University English Linguistics, FR 4.3 English, American and Anglophone Cultures, Campus C5.3 66123 Saarbruecken, Germany Johan Elsness University of Oslo Department of Literature, Area Studies and European Languages Postboks 1003 Blindern 0315 Oslo, Norway Caroline Gentens University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Lobke Ghesquière University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Hilde Hasselgård University of Oslo Department of Literature, Area Studies and European Languages PO box 1003 0315 Oslo, Norway Marije van Hattum Liverpool Hope University Department of English Hope Park L16 9JD Liverpool, United Kingdom Mikko Höglund University of Tampere School of Language, Translation and Literary Studies Kalevantie 4 33014 Tampereen yliopisto, Finland
- 15. Corpus Interrogation and Grammatical Patterns Ditte Kimps University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Georg Maier Allmendweg 4 D-79286 Glottertal, Germany Manfred Markus University of Innsbruck Institüt für Anglistik Knappenweg 9 A 6020 Innsbruck, Austria Nikki van de Pol University of Leuven Department of Linguistics Blijde-Inkomststraat 21 B-3000 Leuven, Belgium Paul Rickman University of Tampere School of Language, Translation and Literary Studies Kalevantie 4 33014 Tampereen yliopisto, Finland Günter Rohdenburg University of Paderborn Department of English and American Studies Warburger Weg 100 33098 Paderborn, Germany Patricia Ronan University of Lausanne Department of English Rychenbergstr. 381 CH 8404 Winterthur, Switzerland Juhani Rudanko University of Tampere Department of English, School of Language, Translation and Literary Studies Kalevantie 4 33014 Tampereen yliopisto, Finland Lieven Vandelanotte University of Namur Unité d’anglais Rue de Bruxelles 61 B-5000 Namur, Belgium
- 16. Interrogating corpora to describe grammatical patterns Kristin Davidse, *Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière KU Leuven (University of Leuven) and *University of Namur Corpus linguistics achieved its first major successes by drawing new types of in- formation about lexical items from large scale corpora. This was the result of in- novative thinking, implemented into creative corpus interrogation tools. From the 1970s on Firth’s (1957: 196) insight that the meaning of a word is defined by its collocates was developed into a corpus methodology (e.g. Jones & Sinclair 1973; Sinclair 1991) that was to revolutionize the field of lexicosemantics. One of its first landmark achievements was the Collins COBUILD English Language Dictionary (Sinclair 1987), a dictionary created from scratch on the basis of sys- tematic study of concordances for each entry. For each lexical item, this allowed the identification of its major subsenses, their most important collocate ranges, their semantic prosody, their near-synonyms and antonyms, and the grammati- cal, or colligational, contexts they occurred in. All these dimensions of lexicose- mantic corpus research are currently being developed further not only in corpus linguistics (e.g. Stubbs 1995, 1996), but also in cognitive linguistics (e.g. Barlow & Kemmer 1994; Tummers, Heylen & Geeraerts 2005) and more formal approaches (e.g. Fellbaum 2004; Fellbaum et al. 2005), and in cognitive psychology (e.g. De Deyne & Storms 2008). The collocational view of lexical meaning is, for instance, informing advanced work on semantic categorization, in which partial synonymy is established on the basis of similarity between collocate clouds (De Deyne, Peirsman & Storms 2009). A similar convergence between theoretical innovation and new corpus methodologies occurred from the 1990s on, when the idea that lexicon and gram- mar form a continuum (e.g. Sinclair 1991; Goldberg 1995, 2006) was translated into specific corpus methodologies (e.g. Hunston & Francis 2000; Stefanowitsch & Gries 2003). What these efforts have in common is that they make “the study of grammar more similar to the study of the lexicon, [...] and more amenable to investigation by
- 17. Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière corpus linguistic methods” (Stefanowitsch & Gries 2003: 210). They produced tech- niques to extract variable idioms and partially filled constructions from large-scale corpora, or to reveal the degree to which particular slots in a construction attract or repel particular lexical items. As Stubbs (2009: 29) put it, “corpus studies have led to an elegant concept of units of meaning which brings lexis in from the cold by inte- grating it within a classic model of linguistic organization”. This volume focuses more on the corpus-based study of grammatical phe- nomena per se. While committed to ‘digging up’ the units of grammar from cor- pus data, the contributions assume a “classic model of linguistic organization” (Stubbs 2009: 29) which investigates selection restrictions on the slots of syn- tagms and looks at which syntagmatic patterns are shared by verbs, nouns and adjectives (e.g. Francis, Hunston & Manning 1996, 1998). The volume documents what practices of grammatical corpus description have been developed by lin- guists who think about grammar ‘with’ corpora. The contributions, which all fo- cus on English, illustrate what mileage can be got out of corpora for advances in grammatical description and argumentation. They show that advances depend on three main conditions: i. guaranteeing optimal representativeness of the corpus data; ii. recognizing essential dimensions of variation between and within corpora; iii. formulating definitions of constructions, and of parameters of the form and semantics of constructions, in such a way that they can be operationalized in the extraction and analysis of datasets. These three principles are embodied in different and often combined ways in the contributions to this volume, as illustrated below. (i) Many of the descriptive advances made by studies in this volume were made possible by the interrogation of corpora which have only recently become available, allowing contributors to fill in gaps in existing documentation across expanded time depth or taking in broader social or regional variation than had hitherto been possible for the topics under investigation. Thus, for instance, El- sness takes a broad “longitudinal look” at the relative frequencies of the present perfect and the preterite in British and American English which previously had been studied only in more limited data sets. Elsness’ contribution makes the im- portant point that corpus interrogation is not just about querying a corpus and then analysing query results taken at face value. The degree to which corpora are truly comparable – even when compiled according to the same design criteria – requires very careful consideration. Initial results culled from the Brown quartet of corpora seemed to suggest a resurgence of the present perfect in the second half of the 20th century – apparently countering a trend established in earlier research. Closer scrutiny, however, revealed large-scale shifts in these corpus texts, with
- 18. Interrogating corpora to describe grammatical patterns more present-time reference overall, especially in going from the corpus repre- senting the American English of the early sixties (Brown) to that representing the early nineties (Frown). Trusting the corpus, then, is not sufficient as a guiding principle when it comes to comparison between corpora, but requires further checking and, where relevant, re-balancing to guard against distortions resulting from cross-corpus imbalances. Markus grapples with the problem of the limited attestation of the ‘a-verbing’ construction, a crucial link in the emergence of the progressive, in the available corpora of written Middle and Early Modern English. He proposes that this gap can be filled somewhat by extrapolation from Wright’s English Dialect Dictionary, now available as the EDD Online, which documents colloquial language across the main English dialects in the United Kingdom in Late Modern English. Dialects being conserving varieties, they can be assumed to throw light on language use in earlier periods. Diemer shows that it may be rele- vant to examine the historical manuscripts themselves, rather than their digitized versions. He is concerned with the development of verb prefixes to prepositional (and adverbial) forms. Manuscripts reveal not only varying spelling practices but also doubt in the mind of scribes, as in periods of transition prefixes would be written as semi-separate forms. Other studies in this volume capitalize on corpora representing regional, temporal and social variation. Comparison of regional vari- ation is found in Blanco-Suárez’ study of the intensifier dead in American, British, Irish and Scottish English, and in van Hattum’s contribution, which compares the uses of can and be able to in nineteenth century Irish English with contemporane- ous British English usage. In Kimps et al’s study of tag questions the different tem- poral provenance and age of the speakers in the London-Lund Corpus of Spoken English and the Bergen Corpus of London Teenage Language provide the occasion for reflecting on the question whether the different frequencies of some subtypes reflect apparent time or real time change. (ii) Grammatical description is also advanced in studies that investigate the impact of text types on grammatical phenomena, treading into – and going be- yond – the tradition pioneered by Biber et al (1999). Often, the study of text types iscombinedwithahistoricalperspective.ThemainaimofvandePolandCuyckens’ diachronic study of the English absolute is precisely to reveal the diffusion of ab- solutes across registers. They argue that the distribution of absolutes has shifted from a system operating along the formal vs. informal cline, and to a lesser extent, the narrative vs. non-narrative cline, to a basic register distinction operating along the literary vs. non-literary cline. Text type also plays a role in Hasselgård’s com- parison of cleft use in writing by – novice and specialist – native speakers of Eng- lish and by Norwegian learners of English. This comparison hinges on the distinc- tion between general argumentative writing and discipline-specific writing. Hasselgård shows that an evaluation of the learners’ use of clefts has to take this
- 19. Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière register difference into account. The learners are found to generally underuse clefts in their English writing, but the picture is more diffuse when specific contexts of use are taken into account. For instance, they overuse clefts in interrogatives in argumentative writing, but underuse clefts with the function of reporting previous research in discipline-specific writing. In Elsness’ study, which examines the changing frequencies of the preterite and present perfect in American and British English, the difference between fiction vs. news-reporting is correlated with differ- ent tense distributions. The preterite is the typical narrative tense of fiction, while newspaper texts are more often oriented towards the present time sphere, and therefore attract the present perfect. Elsness thus draws attention to more abstract features of text types, such as general temporal orientation, which may affect gram- matical choices such as tense. (iii) Finally, advances in grammatical corpus linguistics crucially depend on setting up definitions of grammatical constructions, and of parameters of their form and semantics, in such a way that they can be operationalized in corpus searches and analyses of data. Ronan’s contribution on light verb constructions shows how important a coherent constructional definition is that nets all instances in the various stages of its history. Ronan starts by defining the ‘syntactic context’ of the construction. It is formed by an inflectable verb with low semantic content and a noun that specifies the verbal meaning further, irrespective of whether it is actu- ally derived from a verb (e.g. have a look) or not (e.g. make an effort). As the crucial recognition criterion she takes the possibility to replace the verb – noun expression by a semantically similar simple verb, irrespective of whether it is morphologically related or not to the noun. That is, she applies a functional criterion: does the com- posite phrase serve the same function in clause structure as a simple verb? This provides a coherent definition of the light verb construction as a form-meaning pairing, whose instances range from fully idiomatic to freely co-occurring, in ac- cordance with the current position which holds that constructions may manifest different degrees of compositionality (Goldberg & Jackendoff 2004; Fried 2010). It is precisely this schematic definition that allows her to arrive at a comprehensive view of the quantitative instantiation of the construction in different time periods and correct earlier frequency assessments that were based on a definition in which the possible light verbs were filled in from a Modern English perspective. Interestingly, a number of papers in this volume revisit long-standing prob- lems in syntactic analysis which are given a new spin from a corpus-based per- spective. They involve the study of variant or related constructions, which, rather than being reflected on in terms of invented, decontextualized examples, are stud- ied with the selection restrictions and textual features they manifest in corpora. Höglund revisits the well-known opposition between Mary is eager to please and Mary is easy to please (e.g. Postal 1971, 1974; Jackendoff 1975), where Mary is the
- 20. Interrogating corpora to describe grammatical patterns understood subject of please in the first sentence, which is a control construction, but the understood object of please in the second one, which is a tough construc- tion. He focuses on the case of ready, which untypically allows the formation of both constructions and can in principle lead to instances ambiguous between the two, as in the chicken is ready to eat. He points out that the tough construction can be investigated only if its passive variant which avoids the ambiguity, e.g. the policeman stood ready to be questioned, is brought into the picture. He then exam- ines the role of semantic and textual features in the choice between the active and passive variant, as manifested in historical data slices from Late Modern to Pres- ent-day English. Rudanko and Rickman re-examine from a corpus-based perspec- tive another issue framed within formal grammar. They look at the sentential complements of warn, a verb of the influence type (Sag & Pollard 1991), i.e. one that describes a participant influencing another participant to perform the action in the sentential complement, e.g. I would warn her against paying exorbitant pric- es for books and objects of art. These are control constructions, as the direct object of the matrix sentence is the controller of PRO in the sentential complement. Ac- cording to Bach’s Generalization (Rizzi 1986), the object NP must be structurally represented in object control structures. However, the authors show that warn is attested in covert object control constructions. In a corpus study that looks at the last two centuries, they investigate whether the covert NP object receives a general or a specific interpretation, how it relates to register, and how this developed over time. Rudanko also revisits the hairy issue of the different meanings of the to in- finitival and of -ing complement of the adjective afraid. He singles out Wierzbicka’s (1988) insightful treatment, which associates “tentative intention” with the to in- finitive, e.g. she was afraid to wake her mistress up, and the dreading of what “might happen” with the of -ing form, e.g. he was afraid of going insane. He investigates the comparable semantic contrast between [+choice] for the to infinitive and [–choice] for the of -ing complement in relation to authentic corpus data. Maier studies the distribution of pronoun case forms in it-clefts and it BE sentences, environments in which the occurrence of subject pronouns besides object pronouns has long puzzled scholars. He proposes the hypothesis that the subject pronouns have been reanalysed as postverbal focus markers through markedness reversal, and he links the empirical verification of this hypothesis in corpus data to different variables, set up as necessary and sufficient conditions. Rohdenburg’s study of dual form intensifiers such as real – really starts from the hypothesis that their distribution is parallel to that of simple comparative forms and ones with more, e.g. realer – more real. This hypothesis is confirmed: the simpler forms favour attributive adjectives and non-complemented adjectives while the heavier ones favour non-attributive and complemented adjectives. His datasets allow him to assess whether this distri- bution is prosodically, formally or semantically motivated.
- 21. Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière The fourteen studies included in this volume are ordered in terms of which general grammatical environment they study patterns in: 1. the verb phrase 2. the noun phrase 3. complementation structures 4. clause combining. In Part 1, ‘Patterns in the verb phrase’, the main themes are the internal composi- tion of the verb, tense, aspect, and modality. Patricia Ronan’s contribution deals with light verb, or verbo-nominal, constructions such as make (a) question, take victory, which are generally considered to add aspectual meanings to the strict denotational sense. She argues that diachronic studies so far have too narrowly identified light verb constructions from the perspective of their Modern English patterns with light verbs such as take, have, make, which led to the claim that they are very rare in Old and Middle English. Against this, she has collected all instances of semantically low-content verbs (e.g. niman ‘take’, wyrcan ‘work’, be- ran ‘bear’) followed by action/verbal nouns and forming a single semantic unit in her Old, Middle and Early Modern English corpora. She finds that, on this recon- sidered definition (which is relevant to Present-day English as well), light verb constructions are by no means rare in Old English, and relatively most frequent in the Middle English data. Stefan Diemer revisits the historical shift in English verb-particle combinations from prefixed to prepositional and adverbial forms on the basis of qualitative and quantitative analysis of data from the Helsinki and Wycliffe corpora. He argues that the reasons for the disappearance of the English prefix are more complex than previously thought. He proposes a combination of competition-based and systemic explanations while allowing for additional in- fluence by other developments, such as verb frequency and spelling habits. He also shows that the development is not irreversible, since due to the influence of computer-mediated communication a minor revival of prefix verbs can be ob- served. Manfred Markus’s contribution deals with progressives of the form to be/go a-hunting, which, because of limited attestation in standard corpora, has remained understudied as to its real extent and significance. By extrapolating from Wright’s English Dialect Dictionary, he is able to propose a new take on the historical development of the progressive. The old gerundial construction to be a-verbing did not only play a temporary role as an alternative option side by side with the progressive form, but was, in Late Middle English and Early Modern English, one of the main reasons for the success of the progressive form. He also sheds light on the differential historical spread of the progressive proper. The gerundial construction survived longer in those areas of England where the ger- und collapsed with the ing(e) participle and its dialectal -inde and -ende versions
- 22. Interrogating corpora to describe grammatical patterns because of phonological similarity, viz. in the South and Midlands as well as Scotland and Ireland. The progressive proper thrived more in the North of England, where the participle suffix -and(e) was less affected by the gerundial competitor. Johan Elsness offers a longitudinal look at how the present perfect has developed in competition with the preterite from Modern English on. He shows that the semantic and functional distinction between these two verb forms is not always clear-cut. From Early Modern English on till the present day, the present perfect has been used in combination with specifications of past time such as yesterday morning, last Sunday in breach of the assumed rule barring such combinations. This leaves considerable scope for variation and ambiguity of interpretation. Yet, most of the evidence shows that the present perfect has been in decline from late Modern English on into Present-day English. Marije van Hattum focuses on the specific puzzle of the distribution and frequency of can and be able to in Irish English in the nineteenth century. She shows that in the second half of the 19th century be able to expressing participant-internal possi- bility/ability increased in comparison to both English English in the same period and Irish English in earlier and later periods. She argues that this is due to the fact that some native Irish-speaking writers of 19th century Irish English failed to acquire the subtle differences between these two modal constructions as a re- sult of ‘imperfect learning’. Part 2, ‘Patterns in the noun phrase’, deals with intensification and focus mark- ing. Günter Rohdenburg’s paper explores two novel delicate syntactic constraints on the distribution of suffixless versus suffixed (-ly) intensifiers, which is paralleled by the choice between synthetic and analytic comparatives. He finds that in Present- day corpus data of American and British English, the more explicit vari- ants, analytic comparatives and suffixed intensifiers, are attracted to predicative and other non-attributive adjectives, while attributive adjectives favour the less explicit variants, synthetic comparatives and suffixless intensifiers. For the con- trastive environments of non-complemented and complemented adjectives, the results are not conclusive for non-suffixed versus -ly intensifiers. He considers ex- planations of this distribution in terms of the Complexity Principle, the Principle of Rhythmic Alternation, and the verbality cline on which attributive, postnomi- nal and predicative adjectives can be ranged, and finds verbality to be the main factor. Zeltia Blanco-Suárez provides a detailed discussion of the distribution of the intensifier dead in Contemporary English, with Scottish, Irish, British, and AmericanEnglishbeingsystematicallycompared.Thisintensifierhasbeenclaimed to have been superseded by other intensifiers such as pure in Glasgow and so in Tyneside. It is found to be actually holding out well in Scottish English, but it is most frequent in Irish English, while it is less common in British and American English. Georg Maier focuses on the variation between subjective and objective
- 23. Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière pronoun forms in subject predicatives and the focal constituent of it-clefts. Based on more than 5,000 tokens from the BNC and the COCA, he shows that the distri- bution of pronoun case forms is influenced by several factors such as the clause type, the number and person of the pronoun as well as the mode of discourse. He also demonstrates that the factor Focus, i.e. the relatively most important or salient information in a clause, plays a crucial role. This challenges many earlier accounts of this case of pronominal variation, and supports Maier’s proposal that the sub- ject pronouns have been reanalysed as focus markers. In Part 3, ‘Patterns in complementation structures’, some elusive puzzles of complementation syntax are approached from a novel corpus-based perspective. Juhani Rudanko & Paul Rickman examine the occurrence of covert NP objects in object control structures with the matrix verb warn in written American English of the last two centuries, sourced from COHA. They establish that the covert NP object continues to be rarely specific against claims to the contrary. Its typically vague and indeterminate reference may have fostered the spread of the construction in political English. Juhani Rudanko investigates the variation be- tween to infinitival and of -ing complements of the adjective afraid, through quantitative and qualitative analysis of exhaustive extractions of afraid + infini- tival and of -ing complements from the first three decades of the TIME Magazine corpus. These variants have been explained by linguists such as Bolinger and Wierzbicka in terms of the semantic difference between ‘fear actually present’ (fear of -ing) and fear to undertake ‘potential action’ (fear + to infinitive). He ar- gues for an explanation of the distinction in terms of the semantic roles of the lower subject, e.g. agentive versus experiencer, noting that regularities of statisti- cal significance can be observed when these semantic roles are investigated in a systematic way. Mikko Höglund looks at the two types of complementation con- structions of the adjective ready, illustrated by the two possible readings of The chicken is ready to eat: its control reading in which the chicken is agent, and its tough construction reading, in which the chicken is patient. He presents a dia- chronic-synchronic case study, based on quotations from the Oxford English Dic- tionary online edition and the COCA. He argues that diachronically the tough reading emerged via the intermediary of a construction in which ready was fol- lowed by a passive infinitive. In the synchronic data he analyses the kinds of con- texts in which ready is used in its supposedly more recent, and minor, use as a tough construction predicate, with, for instance, human subject being preferen- tially used in the passive tough construction. Part 4, ‘Patterns of clause combining’, looks at three somewhat peripheral but intriguing cases of clause combining, it-clefts, absolute constructions and tag questions. Nikki van de Pol & Hubert Cuyckens offer a diachronic genre-ac- count of absolute constructions, non-finite constructions consisting of a predicate
- 24. Interrogating corpora to describe grammatical patterns (typically a participle) and a (pro)nominal subject which in their prototypical form are not overtly syntactically linked to the rest of the sentence and can express the same range of semantic relations with their matrix clause as finite subordinate clauses can. The paper explores how the absolute construction diffused from more formal and/or literary genres via less formal genres into the spoken language. In- terestingly, this pattern of diffusion differs from the more commonly assumed path of development, whereby language items spread from the spoken to the writ- ten language. Hilde Hasselgård compares it-cleft use in writing by native speakers of English and by Norwegian learners of English. Detailed qualitative and quantitative comparison is informed by the parameters of general frequency, type of clefted constituent and choice of subordinator in the cleft clause, syntactic envi- ronments of clefts and their discourse functions. Ditte Kimps, Kristin Davidse & Bert Cornillie propose a novel comprehensive description of the speech functions of tag questions: besides real questions they realize statement-question blends, pure statements, commands and offers, and even responses to a preceding ques- tion or statement, which goes against the traditional image of tag questions in reference grammars as questions conducive to answers. The argumentation for the typology integrates formal properties, viz. clause type of the anchor, modals, po- larity patterns, and, in particular, prosody, with conversation analysis (CA), the commodity – information or desired action – being exchanged, and position in turn and in adjacency pair. The different speech function types can be distinguished from each other on the basis of clusters of distinct statistical tenden- cies: each speech function favours some specific values within certain formal and CA parameters. At the beginning of this introduction, the question was raised by what forms of corpus study the description of grammatical phenomena can be advanced. This collection of papers shows many different dimensions of corpus interrogation at work: synchronic -- diachronic, and qualitative – quantitative. Synchronic datasets are interrogated in various ways to re-think specific aspects of grammatical con- structions or to propose new syntheses of the various subtypes of a construction. Historical corpora are queried to arrive at new reconstructions of diachronic stag- es and developments of grammatical constructions. The qualitative analyses are based on constant shunting between data-observation and descriptive generaliza- tions about them, and the corpus data are also interrogated by quantification of observable (sub)patterns and of correlations between features. Between them, the studies illustrate the possibilities of grounding grammatical description in the em- pirical detail that only corpus linguistics can provide, and stimulate researchers to further develop this potential.
- 25. Kristin Davidse, Lieven Vandelanotte, Caroline Gentens and Lobke Ghesquière References Barlow, M. & Kemmer, S. 2000. Usage-Based Models of Language. Stanford CA: CSLI. Biber, D., Johansson, S., Leech, G., Conrad, S. & Finegan, E. 1999. Longman Grammar of Spoken and Written English. London: Longman. De Deyne, S. & Storms, G. 2008. Word associations: Network and semantic properties. Behavior Research Methods 40: 213–231. De Deyne, S., Peirsman, Y. & Storms, G. 2009. Sources of semantic similarity. In Proceedings of the 31st Annual Conference of the Cognitive Science Society, N. Taatgen & H. Van Rijn, H. (eds), 1834–1839. Austin TX: Cognitive Science Society. Fellbaum, C. 2004. Idiome in einem digitalen lexikalischen System. Linguistik und Literatur 34: 56–71. Fellbaum, C., Delfs, L., Wolff, S. & Palmer, M. 2005. Word meaning in dictionaries, corpora, and the speaker’s mind. In Meaningful Texts: The Extraction of Semantic Information from Monolingual and Multilingual Corpora, G. Barnbrook, P., Danielsson & M. Mahlberg (eds), 31–38. Birmingham: Birmingham University Press. Firth, J.R. 1957. Papers in Linguistics 1934–1951. London: OUP. Francis, G., Hunston, S & Manning, E. 1996. Collins COBUILD Grammar Patterns, 1: Verbs. London: HarperCollins. Francis, G., Hunston, S & Manning, E. 1998. Collins COBUILD Grammar Patterns, 2: Nouns and Adjectives. London: HarperCollins. Fried, M. 2010. Grammaticalization and lexicalization effects in participial morphology: A Con- struction Grammar approach to language change. In Formal Evidence in Grammaticaliza- tion Research [Typological Studies in Language 97], A. Van linden, J.C. Verstraete & K. Davidse (eds), 191–223. Amsterdam: John Benjamins. Goldberg, A. 1995. Constructions: A Construction Grammar Approach to Argument Structure. Chicago IL: University of Chicago Press Goldberg, A. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford: OUP. Goldberg, A. & Jackendoff, R. 2004. The English resultative as a family of constructions. Lan- guage 80: 532–568. Hunston, S. & Francis, G. 2000. Pattern Grammar: A Corpus-driven Approach to the Lexical Grammar of English [Studies in Corpus Linguistics 4]. Amsterdam: John Benjamins. Jackendoff, R. 1975. Tough and the trace theory of movement rules. Linguistic Inquiry 6: 437–447. Jones, S. & Sinclair, J. 1973. English lexical collocations. Cahiers de Lexicologie 24: 15–61. Postal, P.M. 1971. Cross-Over Phenomena. New York NY: Holt, Rinehart and Winston. Postal, P.M. 1974. On Raising. Cambridge MA: The MIT Press. Rizzi, L. 1986. Null objects in Italian and the theory of pro. Linguistic Inquiry 17: 501–57. Sag, I. & Pollard, C. 1991. An integrated theory of complement control. Language 67: 63–113. Sinclair, J. et al. (eds). 1987. Collins COBUILD English Language Dictionary. London: Collins. Sinclair, J. 1991. Corpus, Concordance and Collocation. Oxford: OUP. Stefanowitsch, A. & Gries, S. 2003. Collostructions: Investigating the interaction of words and constructions. International Journal of Corpus Linguistics 8: 209–243. Stubbs, M. 1995. Collocations and semantic profile: On the cause of the trouble with quantita- tive studies. Functions of Language 2: 1–33.
- 26. Interrogating corpora to describe grammatical patterns Stubbs, M. 1996. Text and Corpus Analysis: Corpus-Assisted Studies of Language and Culture. Oxford: Blackwell. Stubbs, M. 2009. Technology and phraseology: With notes on the history of corpus linguistics. In Exploring the Lexis-Grammar Interface [Studies in Corpus Linguistics 35], U. Römer & R. Schulze (eds), 15–32. Amsterdam: John Benjamins. Tummers, J., Heylen, K. & Geeraerts, D. 2005. Usage-based approaches in Cognitive Linguistics: A technical state of the art. Corpus Linguistics and Linguistic Theory 1: 225–261. Wierzbicka. A. 1988. The Semantics of Grammar [Studies in Language Companion Series 18]. Amsterdam: John Benjamins.
- 28. part 1 Patterns in the verb phrase
- 30. Light verb constructions in the history of English Patricia Ronan Université de Lausanne This study investigates light verb constructions in sample corpora from Old- Middle- and Early Modern English. The use of one coherent definition of light verb constructions throughout these periods allows direct comparison of the overall structures and of the light verbs used. The comparison shows that frequencies are highest in the Middle English texts and decrease in the Early Modern data. While the Old English counts are significantly lower than Middle English ones, their frequencies are far from negligible. It is argued that where previous assessments consider Old English light verb constructions to be rare or non-existent, this is partly due to having used the perspective of the most frequent Modern English light verbs rather than working from the perspective of which light verbs were frequent at the period in question. Keywords: Old English, Middle English, Early Modern English, analyticisation, language contact 1. Introduction In the present paper, attestations of light verb constructions are compared in sample corpora from Old- to Early Modern English in order to assess changes in the light verbs used. The study intends to offer a survey of historical develop- ments in the frequencies of light verb constructions and to offer a basis for further research based on more extensive data sets. Light verb constructions are defined here as semantically largely non-compositional collocations containing a predi- cate noun which is the syntactic object of a semantically low-content inflected verb, such as to have a look or to take a decision. Typically, but not invariably, these may be paraphrased by simple verbs which correspond to the predicate noun, here to look or to decide. Other terminology used for these constructions includes eventive object, support verb-, stretched verb-, multi verb-, thin verb-, or verbo-nominal constructions.
- 31. Patricia Ronan Concerning the historical development of these structures, it is well known that they already existed in the earliest varieties of English, but the extent of their use is not yet clear. Traugott considers them to be at the very beginning of gram- maticalisation at a ‘phrasal stage’ (1999: 252), in which some fixed collocations exist. In the Middle English period, then, Hiltunen sees them as expanding through French influence (Hiltunen 1983: 28). Previous corpus based studies (especially Brinton and Akimoto 1999) have traced features of the development of light verb constructions throughout the history of English, but the different contributors in- vestigate only the most frequent verbs, and also they partly select the possible predicate nouns according to differing principles. In the following we will suggest a re-evaluation of the relative presence of light verb constructions in different pe- riods of the English language on the basis of an illustrative sample corpus. To this end, we will first summarize previous research on definition and usage of light verb constructions in Sections 1.1 and 1.2, and on light verbs in the history of English in Section 2. Then, after presenting data and methodology in Section 3, we will investigate and compare attestations in our Old English, Middle English and Early Modern English data in Section 4 and discuss our findings in Section 5.1 1.1 Definition of light verb constructions Both synchronic and diachronic comparison of light verbs in English are influ- enced by the fact that the structures do not have a common definition adhered to by all researchers. Consequently, different scholars have also investigated differ- ing entities, which has made particularly diachronic comparison of research results difficult. The most restrictive approach to light verb constructions is to investigate only collocations with zero-derived verbal nominalisations. Nickel (1968) has used this approach in investigating the predicate nouns complement- ing the verbs give, have, make and take, Wierzbicka (1982) restricts herself to predicates of have and take. A less restrictive approach is found in Quirk et al. (1985: 750–2), Algeo (1995: 206) Claridge (2000), and Allerton (2002). These re- searchers include structures that are “semantically an extension of the verb” (Quirk et al. 1985: 750) such as to have an argument versus to argue. As Quirk et al. (1985: ibid.) point out, it is the predicate noun that carries the major part of the meaning of the collocation. Syntactically speaking, the predicate noun is the syntactic object of a semantically low-content inflected verb. Typically, light verb constructions can be paraphrased by simple verbs, for example to have a look by to look or to take a decision by to decide. Quirk et al. (1985: 751) argue, however, 1. I would like to thank the two anonymous reviewers for their useful comments on this study. All remaining shortcomings are of course my own.
- 32. Light verb constructions in the history of English that not all predicate nouns are extensions of simple verbs and cite to make an effort, which is not paralleled by *to effort or to do homework, for which there is no corresponding *to homework. The verbs investigated by Quirk et al. are do, get, give, have, make, offer, pay, put and take. The broader definition of light verb construction used by Quirk et al. is taken as a basis for the research also in this work for the reason that the collocations with action noun predicates and non-action noun predicates are used in similar syntac- tic contexts. In contrast to the studies by the aforementioned authors, also a broader spectrum of verbs is admitted here. The selection criterion is that the en- tire collocation can be paraphrased by a morphologically related or unrelated simple verb at its time of attestation, or will develop a morphologically corre- sponding simplex at a later stage. For these verbal groups, the term constructions has been used from early on. Since the rise of Construction Grammar, this term may be less appropriate. The group verbs at times have form-meaning pairs which are not predictable from their individual components, that is they are semantically non-compositional, but for the moment it remains unanswered to what extent they can be regarded as stored parings of form and function. They are, however, constructions in the sense of denoting a position on a cline between fully idiomatic and complete free co- occurrence, a definition of construction which is used in Goldberg and Jackendoff (2004). According to this view, all related grammatical phenomena, whether idio- syncratic form-meaning pairs or totally general, should be considered together. This view is adopted in the current study. 1.2 Use of light verb constructions and their development The use of light verb constructions is seen as determined by a number of syntactic and pragmatic considerations. It has been pointed out that light verb construc- tions allow for pragmatic reordering, which could not be achieved with simple verbs (e.g. Brinton 1996). (1) a. We dined. b. We had dinner. Sentence (1a) lacks an object noun and the verb dine is in the stress position of the sentence. The light verb construction in sentence (1b), by contrast, provides an object noun for a transitive verb and it also allows a noun to fill the final stress position. The fact that the verbal group consists of two elements entails more syntactic flexibility that can be exploited for pragmatic benefits. In addition, the verbal group can also be semantically more specific. Especially aspectual specifica- tion can be marked in light verb constructions:
- 33. Patricia Ronan (2) a. We walked. b. We took a walk. In example (2a), the action of walking is not temporally restricted, in (2b) the action is temporarily bounded and is potentially repeatable (cf. Wierzbicka 1982: 757–8). In addition to a view from syntactic and semantic perspectives, the use of light verb constructions has also been discussed as a potential language contact or ac- quisition feature. Here the use of light verb constructions has been thought to be favoured particularly in language contact varieties or in children’s speech (Poussa 1990; Tieken Boon van Ostade1990). In language contact, the constructions have been argued to serve as vehicles for the incorporation of new concepts from con- tact languages into a target language (Danchev 1992), and it has been pointed out that this is particularly possible with the verb do (Hock & Joseph 1996). Perhaps due to this association of do with contact varieties, light verb constructions with do have also been identified as signs of colloquial speech, particularly so for Early Modern English (Nurmi 1999). On the other hand, the structures can also be viewed as not stylistically low, but as symptoms of unduly complicated language, as in (3) a. The council has decided to support the project. b. The council has taken a decision to support the project. Brinton (1996: 190) points out that they are often viewed as verbose, scientific or taken as signs of ‘officialese’ or administrative communication. Thus, two different characteristics of light verb constructions, colloquial and elevated style levels, have been argued for so far. 2. Studies on the development of light verb constructions in the history English Light verb constructions have been investigated in the different periods of English, but typically from a synchronic perspective of the period in question rather than from a diachronic perspective. This becomes particularly obvious in Brinton and Akimoto (1999) mentioned above, which is a collection of essays investigating partly differing light verbs in all periods of English language history. Akimoto and Brinton (1999) themselves provide a pioneering study of the structures in Old English. They investigate the Old English equivalents of the light verbs which are most frequent in contemporary English, namely dōn, macian, sellan, giefan, niman and habban. Potential period-specific light verbs are not identified. Ronan (2012a)
- 34. Light verb constructions in the history of English is a corpus based study, drawn from a corpus of about 109,000 words of Old English, which has been collected for cross-linguistic comparison with a morpho- logically even more synthetic language, Old Irish. The study suggests that mor- phological syntheticity is no restraining factor for the use of syntactically analytic light verb constructions. For the Middle English period, Denison (1981) is a comprehensive survey of different kinds of group verbs in one text, the Ormulum, while Matsumoto (1999; 2008) draws on a large corpus in which she identifies uses of the most frequent light verbs. Ronan (2012b) investigates especially the use of foreign-derived pred- icate nouns in some Chaucer materials. For the Early Modern period, Hiltunen (1999) investigates light verbs in drama texts, Tanabe (1999) studies their use in the Paston Letters, while Kytö (1999) examines collocational and idiomatic prop- erties of the most frequent verbs in the combinations. Claridge (2000) is a corpus- based study of different kinds of group verbs in the roughly 1.2 million word Lampeter Corpus, which spans the period from 1640 to 1740 (Claridge 2000: 5). Claridge (2000: 5, 132) identifies 32 verbs which are used in 241 light verb con- structions, with make, take, give, have, put and do leading the table, followed by be, set, lay, beg, call, stand, crave, run, see and bring with more than 10 tokens each (2000: 122). These findings prove that looking beyond the most frequent verbs will provide considerable numbers of light verb types and tokens. Predictably, a number of studies of light verb constructions in Modern English exists, but barring oversight, there are only few comprehensive corpus based stud- ies. Hoffmann (1972) is a corpus based study of light verbs in American drama from the 1950ies and 1960ies. Allerton (2002) examines light verbs in the 1 million word Lancaster-Oslo-Bergin (LOB) corpus, collected in the early 1960ies, but the study is as yet restricted to predicate nouns starting with the letter a. Algeo (1995) has likewise examined LOB, as well as its American counterpart, the Brown cor- pus, for the five most frequent light verbs. In LOB, Algeo found the following counts of light verbs: Table 1. Light verb tokens in the LOB and Brown corpora (after Algeo 1995: 214). Light verb tokens in 1 mil. word LOB tokens in 1 mil. word Brown Tokens per 2 million words have 100 55 155 make 67 59 126 give 40 40 80 Take 38 38 79 do 0 4 4 total 199 245 444 (222/ 1 mil.)
- 35. Patricia Ronan Against this background of studies carried out on contemporary and earlier varieties, let us now move on to consider the developments in data which is sampled on one coherent principle from the Old English to the Early Modern English period. 3. Data and method The data for the present study has been obtained by manual and semi-automatic searches of ca. 50,000 words each of Old English, Middle English, and Early Mod- ern English texts from mixed genres. For the Old English period, data has been collected from the Toronto Dictionary Corpus in Electronic form (diPaulo Healy et al. 2004), and the data is a subset of a ca. 108,800 word data set that contrib- uted to an earlier study (Ronan 2012a). It represents a genre mix of poetry, histor- ical-religious narrative in translation from Latin, original narrative and annalistic evidence (see Table 2). The Middle English data mainly stems from the Helsinki Corpus (Kytö 1996 [1991]) and has been supplemented by data from the General Prologue and The Knights Tale from Chaucer’s Canterbury Tales (de Castro, n.d.). In addition to the latter, original religious narrative has been used in the form of data from The Book of Margery Kempe and short legal material from the Statutes of the Realm. Data from the Early Modern Period has been taken from the Helsinki Corpus only. The intention has been to collect in the corpus texts all instances of semantically low-content verbs that are predicated by predicate noun objects, and in which the resulting verbal group forms a single semantic unit. Thus, the predicate nouns are either action nouns, or abstract nouns whose col- location with verbs could be paralleled by simple verbs, e.g. make a story or do evil. The existence of a corresponding simplex at the period of attestation is not considered a strict criterion, however, as simplexes may have come into existence only later, such as in the case of a/to battle, or, conversely, they may have existed in the past as in the case of to take wife, to which Old English wīfian correspond- ed, but was lost later. Clearly, even if they are not at the time of attestation, light verb constructions may at some point in time be paralleled by a simplex. If any doubt existed as to whether a structure should be considered a light verb con- struction, inclusion has generally been opted for. Overtly aspectual or causal verbs, such as onginnan/gan ‘begin, start’, have however been excluded in this study. These sampling considerations have continuously been followed in all pe- riods under scrutiny. Particularly in more literary genres from the Middle English period onwards, a doubling of predicate nouns was repeatedly found. In order not to distort the
- 36. Light verb constructions in the history of English counts of light verb constructions, here only the predicate noun closer to the verb is used2, e.g. (4) Ne yif us neither mercy ne refuge, ... . (KnT 1720) ‘Give use neither mercy nor escape.’ (5) But swich a cry and swich a wo they make ... . (KnT 900) ‘But they make such cry and such woe.’ In addition, cases were found where two verbs were used together with two predi- cate nouns. In these cases one collocation each was counted, e.g. make corruption and worke disinherison in the following example. (6) Provided also, That no Attainder for this Offence made Felonie by this Acte, shall make or worke any corruption of Blood Losse of Dower or disinherison of Heire or Heires. (Statutes IV 519–20) This methodology has been adopted to make the frequency counts of light verb constructions comparable throughout the periods. As will be seen in Sections 4. and 5, the frequencies of light verb constructions remain high for the Middle Eng- lish period in spite of the count being based on the attestation of verbs rather than predicate nouns. After these methodological preliminaries, we will describe the structures in Old-, Middle and Early Modern English. 4. Data from the history of English 4.1 Old English data The Old English data investigated for this study consists of approximately 45,000 words (cf. Table 2). It predominantly stems from texts from the later Anglo-Saxon period, but the composition date of one text, Beowulf, is still disputed (cf. Orchard 2003: 19–25). In the Old English sample corpus investigated for this study, 93 light verb tokens have been found. In the corpus data, the most frequent verbs do not correspond to contemporary English attestations. Instead, the most frequent verbs were fremman, niman, habban, wyrcan, and dōn and sellan. Particularly interesting is fremman ‘per- form’, which does appear in Bede, but is disproportionately frequent in Beowulf. 2. The author is aware that this can potentially result in incomplete attestation of the predicate nouns in the corpus material. However, as the corpus texts are only a very small selection of the extant material for each of the periods, complete attestation of all light verb constructions with- in the period cannot be aimed for in the context of this study and it seemed most important to focus on comparability of frequency counts.
- 37. Patricia Ronan Table 2. Old English corpus texts. Text Words Genre Period Light verb frequencies/ 100,000 Anglo-Saxon Chron A 15 000 Chronicle 9th-10th century 187 Beowulf 12 000 Poetry 9th – 10th century 333 Ohthere and Wulfstan 2500 Travel narrative 10th century 120 Bede I (extract) 16 500 Historical narrative 10th-11th century 133 Old English total 45000 207 (7) ...swa deorlice dæd gefremede fagum sweordum ‘He performed a valuable deed with shining swords.’ (Beo [0162 (583)]) (8) .. ac ymb Hreosnabeorh eatolne inwitscear oft gefremedon. ‘But often committed dreadful malicious slaughter at Hreosnabeorh.’ (Beo [0681 (2472)]) Fremman seems to correspond to later dōn and macian collocations and denotes active implementation of the action described by the predicate noun. Fremman is used in 15 different types, and particularly help and dæd ‘deed’ show notable token frequencies at 4 and 3 respectively. The predicate nouns may either have an equiva- lent verb or not, though the latter is rarer in the sample. There are no examples of fremman in the samples from Chronicle A or Ohthere and Wulfstan. This may in- dicate that fremman is used more in poetic diction than in non-literary contexts. Semantically, fremman shares some space with wyrcan ‘cause, work’, which is also used in the corpus material. While wyrcan might express causation more than im- plementation in some examples, in others the parallels to fremman are strong, e.g. (9) þær is mid Estum an mægð þæt hi magon cyle gewyrcan ‘The Estonians have the ability that they can create cold’ (Or1 [0148 (1.17.33)]) The parameters that govern this interaction need further research on the basis of a larger corpus. From the sample texts it appears that in addition to semantic con- tent and individual stylistic preferences, alliteration between light verb and predi- cate nouns also favours the use of light verbs constructions. Next in frequency in the data sample is the light verb niman ‘to take’. This too is found with some instances in Bede I, but shows highest token frequency in the Chronicle A, where it particularly appears in apparently set expressions like sige ni- man ‘take victory’ and freod nimman ‘make peace’ (10 examples each). In contrast to fremman, the type frequency for niman in the sample is low, but the tokens appear
- 38. Light verb constructions in the history of English repeatedly, which will be at least partly due to the subject matter of the Chronicles, which are of course much concerned with the topics of victory and peace. (10) þy ilcan geare gefeaht Ecgbryht cyning Beornwulf cyning on Ellendune, Ecgbryht sige nam ‘That same year king Ecgbryht and king Beornwulf fought at Ellendune and Ecgbryht took victory.’ (ChronA [029810 (823.2)]) In spite of its high attestation numbers in these cases, niman may not be entirely grammaticalized in the contexts, as alternative structures exist with habban, as in (11) below. By and large, the collocations with niman indicate or emphasize that the agent benefits from the predicate noun. Habban ‘have’ is a further high-frequency light verb in the sample corpus. Akimoto Brinton (1999: 50) point out that one of the effects of using this light verb with predicate nouns is that using an otherwise obligatory object of an action noun may be avoided, as in cyððe habban ‘to have knowledge’ as opposed to X cyðan ‘make X known’. They further stress that in the nominal construction, the predicate noun may be modified by an adjective, and that this is not possible with the simple verb. The collocation of habban and predicate noun typically expresses a state or a result state as illustrated by (11) hi sona wið heora feondum gefuhtan, sige hæfdan ‘they soon fought with their enemies and had victory.’ (B1 0134 (9.44.28)]. This example illustrates the resultant state of peace. The example is also interesting because sige ‘victory’ has no corresponding simple verb. That this is not of paramount importance for the use of the light verb construction, however, is shown by the fact that its synonym sib, which is paralleled by the simplex sibbian ‘to make peace’, is also used in light verb constructions with habban. A survey of the instances of light verb constructions in the Old English sample corpus is given in Table 3 below. Overall, in the sample texts, the frequency of the light verb constructions was highest in the poetic text of Beowulf with 40 tokens corresponding to a frequency of 333 per 100,000 words. This is followed by the historical Chronicle A, whose 28 tokens correspond to a frequency of 187 per 100,000 words. There is a clear drop in the narrative genres with a frequency of 133 in Bede I (22 tokens) and Ohthere and Wulfstan at a frequency of 120, albeit on the basis of only few tokens (3 tokens). This suggests that genre of the text, stylistically elevated or aiming for exactness of expression, is a more important incentive for the use of the construction in Old English than is the question of whether the text is translated from Latin or not. An overview of the attestations of light verb constructions in the corpus are provided in Table 3.
- 39. Patricia Ronan Table 3. Frequencies of light verb constructions in the Old English sample. Verb Types Tokens Frequency per 100,000 Fremman 15 23 51 Niman 4 22 49 Habban 15 18 40 Wyrcan 8 9 20 Don 7 7 16 Sellan 7 7 16 Limpan 2 3 7 Beran 1 2 4 Dælan 1 1 2 Giefan 1 1 2 Total counts 61 93 207 4.2 Middle English data The Middle English data set for this study consists of approximately 51,000 words taken from different genres. The source texts from the Helsinki Corpus are The Book of Margery Kempe, and the roughly contemporary legal materials from the Statutes of the Realm II from the Helsinki Corpus (Kytö, 1996 [1991]). Two ex- amples of Chaucer’s late 14th century Canterbury Tales, The Prologue and the Knight’s Tale are taken from the eChaucer corpus (Ne Castro, 2007). Texts and data are tabulated in Table 4. In the Middle English sample texts, the line-up of light verbs has changed considerably from the Old English sample. The most frequently used light verbs are have (114), make (61), do (45), take (26), and give (18). Further examples are found with get, go, grant, hold, bring, perform, say, work, arrange, bear, commit, exercise, find, hold, pursue, purvey and put. Compared to the Old English data, both the raw number of tokens, as well as the number of different types of predicate nouns used with the light verbs of high- er frequency has increased. Particularly make, take and give have gained ground, Table 4. Middle English corpus texts. Text Words Genre Period Light verb frequencies/ 100,000 Chaucer 8422 + 21872 Poetry –1380–1400 494 Kempe 9670 Narrative 1436 672 Statutes II 11 271 Legal 1420–1500 790 MidE total 51235 595
- 40. Light verb constructions in the history of English while have and do remain well represented. The most frequent collocation in the samples is have power, of which multiple examples are found in each text, e.g. (12) þat þei schal no powyr haue ouyr þe neyþyr in body ne in sowle; ‘That they shall not have power over you neither in body nor in soul.’(Kempe 151–2) (13) that all such Commyssioners shuld have full power to make ordeigne and execute Statutes. ‘That all such commissioners should have full power to make, ordain and execute Statutes.’ (Statutes II, 25–7) Have collocations are notably better attested than other verbs in the simple narra- tive style of The Book of Margery Kempe. Light verb constructions with do are not very well represented in Kempe nor in the Statutes, but they are found with higher frequency (106/100,000 words) in the Chaucer texts. With the do collocations, predicate nouns which have either a French or a Latin heritage are frequent (cf. Ronan 2012b). Kempe uses 2 of 4 for- eign derived nouns with do, do diligence and do message, but also the English de- rived do business and do evil. In the Statutes II, 5 of 9 predicate nouns are foreign derived. Foreign derived predicate nouns are, however, disproportionally frequent in the Chaucer data (26 of 32 predicate nouns), e.g. (14) And seith “ Arys, and do thyn observaunce.” ‘And says “Arise and do your observance”.’ (KnT 1045) Particularly in rhymed texts, such as the Canterbury Tales, the question of course arises whether the use of light verb constructions must be considered due to a desire to create rhyme. This may hold in some cases, such as do thyn observaunce, which is followed by (15) This maked Emelye have remembraunce To doon honour to May, and for to ryse. ‘This caused Emily do remember to do honour to May, and to rise.’ (KnT 1046–7) In the consecutive eamples (14) and (15), observance and remembraunce with their rhyming French suffix may well have been chosen for rhyming purposes, but the example in the following line, doon honour is clearly not driven by this consider- ation. But here, too, the effect of the light verb construction to increase syllable count as compared to the simplex may have been an incentive (cf. Ronan 2012b). Stylistic considerations may also play a role in legal texts such as the Statutes, which display a strong tendency to consider every angle of a problem (e.g. Stat II, 406–7 below), but in the narrative of Kempe typically few overt stylistic explana- tions of this sort apply.
- 41. Patricia Ronan (16) By which Co[m]myssions, and auctoritie yeven to the seid Co[m]myssioners in the seid fourme, meny greate hurtes and inconvenyences in dyv[er]s p[ar] ties of this yor realme doon and had by encresse of Water were necessarily redressed refourmed and amended... ‘Many great hurts and inconveniences done and had in diverse parts of your realm by increase of water were necessarily redressed, reformed and amended by the commissioners and the authority given to the said com- missioners in the said form.’ (Statutes II 68–72) (17) so sche dede hir massage as sche was comawndyd. ‘And so she did her errand as she was commanded.’ (Kempe 218–19) While in some cases stylistic considerations like rhyme, rhythm or other prag- matic considerations may cause the use of the light verb, in this last example no clear alternative involving a simple equivalent to do a message seems to exist. The new light verbs in this period, make and take ostensibly replace fremman and niman, and give replaces sellan, which had become more restricted in meaning after the OE period. Though it has a more recent history as a light verb than do, make has overtaken it in the sample texts from the Middle English period, where it is already well represented with 38 tokens, though this might look different in a larger corpus. Statutes II particularly use it in the context make feoffment ‘grant a fief’, for which the simplex feoffen also existed (OED, sv feoff). Otherwise, make mention is well attested, and make act and make list are also found repeatedly. (18) þe creatur of whom þis tretys makyth mencyon ‘The creature of whom this treatise makes mention.’ (Kempe 456–7) There are also further examples with rarer verbs which continue to be medium frequent or rare support verbs in contemporary English, such as grant or put (in) (Allerton, 2002, pp. 186, 190), and an example of commit is in evidence: (19) ... where the Attourney named in the seid Warrant put the seid Article in execucion ‘where the attorney named in the said warrant put the said article in execution’ (Statues II 628–9) (20) ... all man~ of accions [...] to graunt to ev~y of such p~sones p~teccion [...] ‘all manner of actions to grant protection to all such persons’ (Statutes II, 292–4) (21) Forasmuche as Thomas Crofte commytted a detestable murdre within the Marches of Wales ‘Forasmuch as Thomas Crofte committed a detestable murder within the Marches of Wales’ (Statutes II, 829)
- 42. Light verb constructions in the history of English Table 5. Light verbs in the Middle English sample. Verb Types Tokens Token f/100,000 Have 75 114 223 Make 38 61 119 Do 32 45 88 Take 18 26 50 Give 14 18 35 Others 25 41 61 Total counts 232 305 595 The distributions of light verbs in the Middle English samples are shown in Table 5. Table 5 shows that the frequency of light verb constructions has risen to 595 examples per 100,000 words in the sample corpus, as compared to 207 examples per 100,000 words in the Old English sample texts. While the increase of almost 200% is indeed very notable, it is not so exorbitant that we could argue that the structure is only emerging in the Old English period. But the higher type frequen- cies in Middle English indicate that the semantic spread of the constructions seems to be increasing. 4.3 Early Modern English data The Early Modern data sample totals about 53,000 words, all taken from the Helsinki Corpus. A genre mix has again been attempted. The eight textual extracts also contain legal material from the Statutes of the Realm IV, rhyming drama as well as non-rhyming drama, historical writing, travelogue and a sermon as indi- cated in Table 6. In the Early Modern material, have and make retain their positions as most frequent light verbs. They are joined by give, do, commit and take as more fre- quently used verbs, and further examples are found with exercise (5), put (5), hold (4), use (4), fight (3), levy (3), practice (3), receive (3) and work (3). Execute, get, grant, keep, minister and perform have 2 attestations each, bring, break, come, draw, gain, lead, obtain, offer, perpetrate, serve, set, stand, undertake and yield have one attestation each in the corpus. The data in Table 7 shows that in comparison with the Middle English sample, the frequencies of light verb constructions are lower in the Early Modern sample texts. The data displays a large diversification in light verbs. At the same time, the five most prominent light verbs are already those generally observed in Modern English. Commit has an equally high token frequency as take. This however, may
- 43. Patricia Ronan Table 6. Early Modern English corpus texts. Text Words Genre Period Light verb frequencies/ 100,000 Statutes IV 11 795 Legal 1588–1604 678 Merry Wives of Windsor – Shakespeare 6245 Comedy – rhyming 1597–1623 464 A chaste maid in Cheapside, Middleton 5708 Comedy – non-rhyming 1630 210 History of Britain Milton 6002 History 1670 450 On the folly of scoffing at religion Tillotsen 6702 Sermon 1671 433 New account of East India and Persia Fryer 5438 Travelouge 1698 221 Reign of Charles II Burnet 5280 History 1703 704 Beaux Strategem Farquhar 5704 Drama 1707 245 Total EModE 52874 463 Table 7. Light verbs in the Early Modern sample. Verb Types Tokens Token f/100.000 Have 42 58 110 Make 39 52 98 Give 25 25 47 Do 18 20 39 Take 11 16 30 Commit 3 16 30 Exercise 4 5 9 Put 3 5 9 Hold 4 4 8 Use 4 4 8 Others 37 41 78 Total counts 186 245 463 be document specific as the high counts are mainly based on the high type fre- quencies of commit offence in the Statutes IV (13 tokens). In the restricted sample corpus, only a small number of light verb types are found repeatedly, and repetition is predictably most prominent in light verbs of
- 44. Light verb constructions in the history of English high frequency. Do and give largely have single attestations with their predicate nouns (excepting do evil, felony and evil at 2 tokens each). The light verbs have, make and take, on the other hand, show a higher number of multiple attestations. This holds for have power, have regard, have authority, have faculty, and for have charme, confidence, means, need and opportunity. Make is found with multiple in- stances of use, discovery, statute, claim, peace, proof, provision and rates. Take is found in the collocations take heed, leave, care and distress. The latter is used in the sense of ‘detention’ in the Statutes, and is paralleled by a simple verb from 1490 on (OED sv distress n II). The following samples illustrate the Early Modern English usage in the data sample: (22) He had a faculty of speaking indefatigably upon every subject (Burnet, Reign 228–9) (23) ... of whose Treason, the King having now made use as much as serv’d his turn (Milton, 425–7) (24) ... and the surplusage or remainder over and above to be delivered to the partie of whome the distresse was taken (Statutes IV 450–1) The question has previously been raised, whether the presence of a definite article in the context of a predicate noun of a light verb means that the collocation cannot be considered a light verb construction. Particularly for earlier stages of the English language, this is contradicted by researchers who point out that the article system was less fixed at earlier stages of English (e.g. Matsumoto 2008: 84 for Middle English). Wierzbicka (1982: 758) finds the presence of an indefinite article to be most typical in Present Day English, and she identifies it as semantically singular- izing and delimiting. If we are prepared to admit indefinite articles in collocations like to take a decision or to have a think, some situations may also call for the use of the definite article, especially where previously mentioned light verb construc- tions are referred to, e.g. have you taken the decision we are waiting for? These two factors, less fixed state of articles especially in pre-Early Modern English, and con- textual anchoring of definiteness, suggest that light verb constructions with defi- nite articles should not be dismissed. In the corpus material from this period it is noteworthy that as indicated above there is considerable variety in the light verbs used, such as (25) And if any Weaver shall use any Deceipte in mynglinge his Yearne (Statutes IV 348) (26) when I absolutely renounce their Authority, I do yield a most perfect sub- mission and obedience to it (Tillotson, 444–5) (27) Therefore when we undertake this Journy (Fryer, 347–8)
- 45. Patricia Ronan This large variety of different light verbs may in part be influenced by the presence of large groups of morphologically differing eventive nouns that have become available at the beginning of the Early Modern period. On the other hand, the fact that denominal verbs could also be formed to a larger extent may have meant that it was often not necessary to use a light verb construction anymore because a sim- plex had become available. This latter possibility may cause frequency decreases in light verb constructions in the Early Modern English period. Overall, in the corpus material especially high frequencies are found in the examples of legal language, a mid-position is taken by rhyming drama, historical and religious writing, the lowest frequencies are presented in the more casual speech genres of non-rhyming drama and travelogue. Burnet’s Reign of Charles II seems an unexpected outlier (704 examples per 100,000 words). Even the Stat- utes IV, earlier than Burnet, have fewer examples (678 per 100,000)3. It is note- worthy that the average frequency is lower in the Early Modern English data compared to the Middle English data, even in examples of identical genres such as statutes. 5. Discussion The corpus material indicates that attestations of light verb constructions rose over time. The rise of frequency was not steady, however. We can observe an increase of light verb constructions, both in terms of frequencies and attested structures in the Middle English data as compared to the Old English data. This rise in frequency from Old to Middle English is not exorbitant: the constructions are about two times more frequent in the Middle English data than in the Old English data. The differences in frequencies of support verb constructions between Old-, Middle- and Early Modern English corpus texts is statistically highly significant at a sig- nificance level of p 0.0001%. This rise is clear, but it does not seem a sufficient basis for claiming that the constructions were badly attested in Old English. The overall distribution of light verb constructions in the corpus material is plotted in Graph 1. The graph illustrates a rise from Old English to Middle English, which is fol- lowed by a fall in the Early Modern English corpus data. For the rise in Middle English frequencies, a number of possible reasons may be found. One important reason for using light verb constructions is to augment the lexicon. Augmentation may prove necessary or desirable because a simple expression did not exist at the 3. The difference between Burnet’s Reign of Charles II and the other EModE texts is statisti- cally significant at a significance level of 0.04% according to the chi-square test.
- 46. Light verb constructions in the history of English 0 O h t h e r e B e d e I C h r o n A B e o w u l f C h a u c e r K e m p e S t a t u t e s I I S t a t u t e s I V M e r r y W i v e s C h a s t e M a i d M i l t o n H i s t o r y T i l l o t s e n S e r m o n F r y e r P e r s i a B u r n e t R e i g n F a r q u h a r B e a u x 100 200 300 400 500 600 700 800 Graph 1. Frequencies of light verb constructions per 100,000 words in the corpus texts. time, for example where technical innovations are concerned, or where concepts are loaned from other languages. Particularly in cases of loaned verbs from other languages, these often are loaned as agent nouns first, and used with a light verb, particularly do, before they are fully integrated into the language as inflected verbs. This is due to the fact that nouns are more easily borrowed from other languages than verbs (cf. Matras 2008: 172). An example of this is I, that wende and hadde a great opinioun... ‘I supposed and had a great notion’ (Chaucer, Knight’s Tale, 1268). Opinion is a loan from French or late Latin, the verb opinion also comes to be used in English, but only after Chaucer’s time (cf. Ronan 2012b). In this respect, light verb construction may be seen as a language contact phenomenon, and it has been argued that they are frequently found in contexts of second language acquisition (Poussa 1990; Rissanen 1991: 333–35; Danchev 1992: 30). As language contact was very strong particularly in the Middle English period, high frequencies of light verb constructions would not have been unexpected at this time. In addition to language contact phenomena playing a role, it has been sug- gested that the decrease in morphological affixation may have favoured the use of analytic structures and thus of light verb constructions (Hiltunen 1983). Since growing analyticisation was particularly observable in the Middle English period, this would also favour the use of light verb constructions in Middle English. In this context, the increasing fixation of the word order of English may also have favoured analytic expressions that allow syntactic and pragmatic modification of the sentence structure. A further point that has not been examined in this study, but that would offer interesting insights into the use of light verb constructions, is the question of how the spread of literacy during the Middle and particularly Early Modern periods
- 47. Patricia Ronan may have influenced the use of light verb constructions. That the sociolinguistic context may have played a role is possible, given that Nurmi (1999) identified par- ticularly do periphrasis as a feature that is prominent in the language of speakers of lower social classes. There is some suggestion that education may have an influ- ence in the fact that the data from The Book of Margery Kempe, who was herself illiterate and uneducated and had a local clergyman write for her, shows a high frequency of light verb constructions, particularly with the light verb have. It is also interesting to note that the data from the Early Modern texts in the corpus do not indicate a further rise in the frequencies of light verb constructions after the Middle English period, except for the outlier of Burnet’s Reign of King Charles. On the one hand, this may be due to data selection and only a consider- ably larger data sample from each period of English can prove the frequency changes. On the other hand, Danchev (1992) has argued that during English lan- guage history, periods of decreased language contact have always resulted in a more synthetic language structure. For the Early Modern English period it can probably be argued that there was comparatively less language contact than during the Middle English period, which could have meant an increasing syntheticity of the verbal paradigm not only in terms of Latinate prefixes, but also in terms of a reduced incentive to use light verb construction. 6. Conclusion This investigation of light verb constructions shows differences from previous corpus-based surveys which predominantly investigated the distribution of high- frequency light verb constructions, which were partly selected on the basis of fre- quencies of light verbs in Present Day English. When using identical selection criteria for light verb constructions, we find that the frequencies of attestations rise considerably in the later Middle English period as compared to Old English. The overall frequencies in the Middle English data are in fact the highest in the corpus material, after the late Middle English period, a slight drop in frequencies is observable in the Early Modern English corpus texts. In contrast to earlier studies the corpus material suggests, however, that the frequencies of light verb constructions in Old English are not insignificant or negligible as the frequencies amount to more than 1/3 of the high Middle English frequencies, or in other words, the Middle English frequencies are almost 200% higher than the frequencies of the Old English corpus data. The mis- apprehension that there were hardly any light verb constructions in Old English at least partly seems to arise from the practice of investigating predominantly those light verbs which are still frequent in Present Day English. The most frequent
- 48. Light verb constructions in the history of English Present Day English light verbs are already well represented in the Middle English corpus data, but in Old English there is only a partial overlap with Present Day English verbs, which means that a large percentage of Old English structures is overlooked in that approach. The present study has also asked the question of why the numbers of attestation may be the highest in Middle English corpus data and it is suggested that language contact may have played an important role in the increase in frequency. However, this assumption, as well as the observed frequen- cies, should be tested on considerably larger corpora comprising more varied cor- pus texts and genres. References Akimoto, M. Brinton, L. 1999. The origin of the composite predicate in Old English. In Col- locational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 21–58. Algeo, J. 1995. Having a look at the expanded predicate. In The Verb in Contemporary English. Theory and Description, B. Aarts C.F. Meyer (eds), 203–217. Cambridge: CUP. Allerton, D. 2002. Stretched Verb Constructions in English. London: Routledge. Brinton, L. 1996. Attitudes to increasing segmentalization. Complex and phrasal verbs in Eng- lish. Journal of English Linguistics 24: 186–205. Brinton, L. M. Akimoto. (eds). 1999. Collocational and Idiomatic Aspects of Composite Predi- cates in the History of English. [Studies in Language Companion Series 47]. Amsterdam: John Benjamins. Claridge, C. 2000. Multi-word Verbs in Early Modern English. Amsterdam: Rodopi. Danchev, A. 1992. The evidence for analytic and synthetic developments in English. In History of Englishes: New Methods and Interpretations in Historical Linguistics, M. Rissanen, O. Ih- alainen, T. Nevalainen I. Taavitsainen (eds), 25–41. Berlin: Mouton de Gruyter. Denison, D. 1981. Aspects of the History of English Group-Verbs: With particular Attention to the Syntax of the Ormulum. PhD dissertation, University of Oxford. diPaulo Healy, A., D. Haines, J. Holland, D. McDougall, I. McDougall X. Xiang (eds). 2004. The Dictionary of Old English Corpus in Electronic Form. TEI-P3 conformant and TEI-P4 conformant version. Toronto: DOE Project. eChaucer: NeCastro, G. 2007. eChaucer.http://www.umm.maine.edu/faculty/necastro/chau- cer/ (29 May 2012) Goldberg, A. Jackendoff, R. 2004. The English resultative as a family of constructions. Lan- guage 80: 532–568. Hiltunen, R. 1983. The Decline of the Prefixes and the Beginnings of the English Phrasal Verb [An- nales Universitatis Turkuensis Series B, Vol. 160]. Turku: University of Turku Press. Hiltunen, R. 1999. Verbal phrases in Early Modern English. In Collocational and Idiomatic As- pects of Composite Predicates in the History of English, Brinton Akimoto (eds), 133–166. Hock, H.H. Joseph, B. 1996. Language History, Language Change, and Language Relationship: An Introduction to Historical and Comparative Linguistics. Berlin: Mouton de Gruyter.
- 49. Patricia Ronan Hoffmann, A. 1972. Die verbo-nominal Konstruktion – eine spezifische Form der nominalen Ausdrucksweise im modernen Englisch. Zeitschrift für Anglistik und Amerikanistik 20: 158–183. Kytö, M. (ed.) [1991]1996. Manual to the Diachronic Part of the Helsinki Corpus of English Texts. Coding Conventions and Lists of Source Texts, 3rd edn. Helsinki: Department of English, University of Helsinki. Kytö, M. 1999. Collocational and idiomatic aspects of verbs in Early Modern English. In Collo- cational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 167–206. Matsumoto, M. 1999. Composite predicates in Middle English. In Collocational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 59–96. Matsumoto, M. 2008. From Simple Verbs to Periphrastic Expressions. Frankfurt: Peter Lang. Nickel, G.1968. Complex verbal structures in English. IRAL - International Review of Applied Linguistics in Language Teaching 6: 1-21. Nurmi, A. 1999. A Social History of Periphrastic DO. Helsinki: Société Néophilologique. Orchard, A. 2003. A Critical Companion to Bēowulf. Cambridge: D.S. Brewer. Poussa, P. 1990. A contact-universals origin for periphrastic do, with special consideration of OE-Celtic contact. In Papers from the 5th Internatioanl Conference on English Historical Linguistics [Current Issues in Linguistic Theory 65], S. Adamson, V. A. Law, N. Vincent S. Wright (eds), 407–434. Amsterdam: John Benjamins. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the English Language. London: Longman. Rissanen, M. 1991. Spoken language and the history of do-periphrasis. In Historical English Syntax, D. Kastovsky (ed.), 321–342. Berlin: Mouton de Gruyter. Ronan, P. 2012a. Make peace and take victory. Support Verb Constructions in Old English in Comparison with Old Irish. NOWELE Supplement Series 24. Odense: University of South- ern Denmark Press. Ronan, P. 2012b. Mobilizing linguistic concepts: support verb structures in Early English. In English on the move: Mobilities in literature and language [SPELL 27], D. Britain A. Kern- Staehler (eds), 183–99. Tübingen: Narr. Tanabe, H. 1999. Composite predicates and phrasal verbs in the Paston Letters. In Collocational and Idiomatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 97–132. Tieken Boon van Ostade, I. 1990. The origin and development of periphrastic auxiliary do: A case of destigmatisation. North-Western European Language Evolution (NOWELE): 16: 3–52. Traugott, E.C. 1999. A historical overview of complex predicate types. In Collocational and Idi- omatic Aspects of Composite Predicates in the History of English, Brinton Akimoto (eds), 239–260. Wierzbicka, A. 1982. Why can you have a drink when you can’t *have an eat? Language 58: 753–99.
- 50. What happened to the English prefix, and could it stage a comeback? Stefan Diemer Saarland University, Germany This paper revisits the historical shift in English verb-particle combinations from prefixed to prepositional and adverbial forms based on qualitative and quantitative examples from the Helsinki and Wycliffe corpora collected during a study on the history of verb-particle combinations (Diemer 2008). It is argued that the reasons for the disappearance of the English prefix are more complex than previously thought. The paper proposes a combination of competition- based and systemic reasons while allowing for additional influence by other developments, such as verb frequency and spelling habits. Excerpts from corpus- based studies (Diemer 2009, 2013) show that the development is not irreversible, since due to the influence of computer-mediated communication there may be a revival of prefix verbs modelled after Old English templates. Keywords: verb-particle combinations, prefix, computer-mediated communication 1. The prefix then and now The Old English (OE) prefixed verb was definitely a success story, as its considerable share of total verb-particle combinations shows. Due to its elaborate affix system, the late Old English verb system was of high syntactical and functional complexity. More than thirty highly productive prefixes existed, accounting for up to a quarter of all verb-particle combinations (as quantified in Diemer 2008). These verb forms exhibited literal, but also figurative meanings and could thus perform all functions of present-day English phrasal verbs. Examples (1) and (2) show these different functions for the Old English verbs ingan (to go in) and ondrædan (to fear). (1) þæs huses þær seo fæmne ineode ‘The houses in which the maid went’ (Anonymous, Martyrology, Helsinki Corpus 1399 OE2)
- 51. Stefan Diemer (2) Sume seoce synd swa dysige, þæt hy ondrædað him ‘Some sick are so ignorant that they dread him’ (Aelfric, First and Second Letter, Helsinki Corpus 36 OE3) However, a thousand years later the language has changed fundamentally. Present- day English only retains a very small percentage of productive prefixes, accounting for less than 3 percent of all verb-particle combinations (Diemer 2008), which to- day are mostly adverbial and prepositional constructions. Why did a combination that is still popular in other Germanic languages to this day disappear rather sud- denly? This article will investigate the reason for this drastic shift and describe possible factors that prompted it. 2. Particles and their variation in early English The term “prefix” describes a morphological category from a structural perspec- tive. Although it has been criticized and avoided by researchers such as Adams (2001) or Plag (2003), mainly because of its lack of semantic precision, it is pre- ferred here because it has two significant advantages: it can easily be found and quantified in a diachronic text corpus, and it can be compared to other morpho- syntactic categories such as prepositional and adverbial verbal compounds. In de- fining prefixes this article follows Hans Marchand, who describes prefixes as “bound morphemes preposed to free morphemes” (Marchand 1969: 129). In or- der to perform this investigation, it is necessary to distinguish two types of pre- fixes: (1) non-lexical and (2) lexical prefixes. (1) Non-lexical prefixes: Originally, all bound morphemes probably were de- tachable, functioning as prepositions or adverbs, but already in OE there are many that were fully integrated and could not stand on their own, like a-, be- etc. Hiltunen (1983: 47) pays special attention to them and concludes that these would be particularly vulnerable to changes in syntax since they could not move. Lexi- cally these were replaced by analytical structures, while stylistically they became integrated with the verb meaning. From a research perspective, these prefixes, by eliminating syntactic factors, are certainly useful in examining lexical and seman- tic factors that contributed to their decline. However, because of this, they also do not allow the quantitative comparison to other morphosyntactic categories. There- fore, this article does not include them. (2) Lexical prefixes: Much more illuminating from a morphosyntactic point of view is the fate of those prefixes that can still function as adverbs or prepositions, such as over-, under-, with-, since their movement assumedly is not restricted by grammatical factors. It is these prefixes that are the focus of attention here; this article will use the term “lexical prefixes” for them, although some researchers
- 52. What happened to the English prefix? such as de la Cruz (1975) use the unfortunate term “separable prefixes”. Marchand (1969: 109) remarks that of those remaining lexeme prefixes, only out-, over- and under- are still productive (apart from remnant forms), excluding, for the most part, verbs with up- (identified as Latin translations or analogues), back- and off- (categorized as conversions or back-formations). Lexical prefixes are used as the basis for this study, since they can exhibit all types of variation. Marchand’s definition of “bound” also needs to be further differentiated, since there are various possible degrees of connection between the two morphemes, particularly in Middle English. A verbal prefix can be 1. joined to the verb (examples 3.1., 3.4 and 3.5) 2. hyphenated (example 3.2.) 3. separated (example 3.3) Prefixes are thus, even on a purely morphosyntactic level, not as straightforward to distinguish as it may seem. A prefix is, in fact, a rather elusive form, which may explain why it is treated so differently in the various corpora. Old and Middle English prefixes could, to provide one salient example, still separate from their respective verbs, whether they were, strictly speaking, lexemes or not. This separa- tion could manifest itself in several ways: Particles could attach to the verb, be hyphenated, stand separately in front of it and “take a stroll”, as it were, exhibiting varying degrees of removal from the main verb. Thus, the phrase He went into that temple could be realized in various ways in Old English (and most of the Middle English period), as example (3) shows. (3) “He went into that temple” Prefixation options in OE a. He ineode þæt templ b. He in-eode þæt templ c. He in eode þæt templ d. He ineode in þæt templ e. He ineode on þæt templ Other possible constructions in OE d. He eode in þæt templ f. He eode þæt templ in g. Eode he þæt templ in The morphosyntactic flexibility of the particle needs to be taken into account by creating at least one intermediate category which I propose to call “separate pre- fixes”. It is, as its diachronic share of total verb-particle combinations will show, a classical intermediate category. This distinction still does not adequately reflect the many shades of remoteness actually found in manuscripts. “Taking a stroll”
- 53. Stefan Diemer seems quite a fanciful phrase, but it is not used figuratively: in the underlying manuscripts, distance becomes indeed a variable, with particles and articles at- taching to and detaching from their respective partners. Examples (4) to (8) illus- trate this variability. The illustrations are from Wycliffite texts written around 1380 AD, thus from the (for prefixation late) period Middle English 3 (ME3, as per the Helsinki Corpus designation). The by far largest share of prefixed verbs follow es- tablished prefix patterns, as shown in example (4) with the verb forsake, the same usage as in Present-day English. (4) forsake A man schal forsake fader [...] ‘A man shall forsake father [...]’ (MS Bodley 296 1v, detail) Example (5) shows the prefixed verb totake, in which the prefix to is written in a slightly different font and at about half the usual distance between words. (5) totake totake marie þi wyf [...] ‘to take Mary your wife [...]’ (MS Junius 29 13v, detail) Example (6) shows the non-lexical prefix a- separated from -shamed, creating a rather unusual group of two non-lexical morphemes forming one discontinuous lexeme. (6) a schamed and þei weren not a schamed [...] ‘and they were not ashamed [...]’ (MS Bodley 296 1v, detail) Example (7) illustrates that even articles could be prefixed, raising the question of whether a morphosyntactic analysis of article prefixation in Middle English could produce qualitative differences between different types of articles. (7) þeday “þei were before to þeday [...]” ‘they were before to the day [...]’ (MS Douce 370 5v detail)
- 54. What happened to the English prefix? Example (8) shows the preposition into written as two separate particles. (8) clepe in to (PDE: call sb.) god: clepe in to adam [...] ‘God called Adam [...]’ (MS Douce 370 6r detail) These manuscripts also illustrate the advantage of a hand-written text: if a scribe was not sure whether to separate the particle or not, he could muddle the waters and semi-separate it, as example (9) shows with two different versions of into. (9) in to (MS Laud misc 361, lv, and MS Bodley 665, 13v, details) Diachronic corpora (such as the Helsinki Corpus), for the most part, do not dis- tinguish these forms; if the data is tagged, these stages of removal are not uni- formly categorized. In view of these taxonomic problems it is not surprising that most research on the morphosyntactic category of prefixes has been qualitative or based on small quantitative samples, which may explain the wide range of answers to the question why the prefix disappeared. It is not easy to discuss the disappear- ance of the prefix on a global scale, as various prefixes may behave differently. Furthermore, a quantitative analysis is difficult for the reasons mentioned above. 3. The decline of the prefix: Theories Broadly speaking, three different hypotheses are proposed to explain the decline of the prefix: 1. lexical change and competition 2. syntactical change 3. semantic change
- 55. Exploring the Variety of Random Documents with Different Content
- 56. G. P. W. Collection. Noronhia emarginata. This tree is a native of Madagascar and also of Mauritius. A fine specimen may be seen at the Government Nursery, Honolulu. It is a handsome evergreen with entire, cuneate, coriacious leaves, having short petioles. The yellowish flowers come in clusters, and are quite fragrant. The fruit is a one-celled drupe, almost round, and about an inch in diameter. It is purple when ripe, and has a tough skin. The sweet, edible pulp surrounds a very large seed. Plate LXXV
- 57. Plate LXXV.—Noronhia emarginata. One half natural size.
- 58. G. P. W. Collection. Castanea sativa. JAPANESE CHESTNUT. This is a close-headed tree of slender growth, attaining a height of from 30 to 50 feet. Its leaves are smaller than those of other chestnuts, generally from 3 to 7 inches long, and are either rounded at the base or reduced to a long, bristle-like point. The monoecious flowers are arranged in long catkins. The small burs have a thin, papery lining, and short, widely-branching spines. The nuts are large and glossy, usually three in a bur. They are somewhat inferior in quality, but are palatable when cooked. To my knowledge there is but one tree of this variety growing in these Islands, and it is to be found on the slopes of Tantalus, where it was planted by the Department of Agriculture. Plate LXXVI
- 59. Plate LXXVI.—Japanese Chestnut. One half natural size.
- 60. G. P. W. Collection. Inocarpus edulis. TAHITIAN CHESTNUT. This tree, which is said to be a native of the Moluccas, is an evergreen of very rapid growth. Its straight trunk, with smooth, ashen-grey bark, its spreading branches, with their dense green foliage, make a very ornamental as well as useful tree. Its leaves are alternate and simple. The small, fragrant, pale yellow flowers are very numerous. The drupe is obliquely oval, and about the size of a goose egg, containing a large kernel which is edible when roasted, but is not especially palatable. The only trees of this variety growing in Hawaii are to be found at Ahuimanu Ranch, Oahu, where they fruit regularly, and the seeds germinate after being in the ground some months. Plate LXXVII
- 61. Plate LXXVII.—Tahitian Chestnut. One third natural size.
- 62. G. P. W. Collection. Canarium commune. CANARY NUT. This medium-sized nut-bearing tree is found growing in Java, Guam and the Philippines, and from any one of those countries may have been introduced to Hawaii. A fine specimen may be seen at the Government Nursery, Honolulu. Its leaves are alternate, odd pinnate. The small flowers come in terminal panicles. The fruit or nut is ellipsoidal. The thick skin, which is purple-colored when ripe, covers a hard, three-lobed stone, which differs from a pecan nut only in that it is sharp at each end. The kernel is small, sweet and edible. Trees propagated from the mature nuts. Plate LXXVIII
- 63. Plate LXXVIII.—Canary Nut. One half natural size.
- 64. G. P. W. Collection. Canarium commune. CANARY NUT (round variety). Few trees of this round variety are to be found in Hawaii. Its leaves are smaller than those of the preceding variety, and it is a very poor bearer. Plate LXXIX
- 65. Plate LXXIX.—Canary Nut (round var.) One half natural size.
- 66. G. P. W. Collection. Macadamia ternifolia. QUEENSLAND NUT. This sub-tropical Australian tree sometimes grows to a height of 60 feet, but in Hawaii is of medium size. It is symmetrical and handsome, having dark green, shiny foliage, and long tassel-like white flowers. Its glabrous leaves are sessile, oblong, lanceolate, serrate, with fine prickly teeth, and come in whorls of 3 to 4, varying in length from a few inches to a foot. Flowers small; fruit has a thick, very hard shell, which when ripe is a smooth, shiny brown. The kernel is white, crisp and sweet, and has the flavor of hazel nuts. It may be eaten either raw or roasted. The tree matures its fruit in the Fall months, and is easily propagated from the fresh nuts. Plate LXXX
- 67. Plate LXXX.—Queensland Nut. One half natural size.
- 68. G. P. W. Collection. Macadamia sp. This variety of the Queensland nut has leaves and fruit larger than those of Macadamia ternifolia. Plate LXXXI
- 69. Plate LXXXI.—Macadamia sp. One half natural size.
- 70. G. P. W. Collection. Aegle Marmelos. BHEL OR BAEL FRUIT. This small spinose tree is a native of tropical Asia, and although not commonly grown in Hawaii, specimens may be found in several gardens. It has alternate trifoliolate leaves, and flowers, which grow in clusters, are small and fragrant. The gourd-like fruit, with its hard shell, is from 2 to 4 inches in diameter, and is either round or pear- shaped, and although heavy and solid, it will float in water. The rind, when ripe, is a yellowish-brown color, and is studded with oil cells. The interior surface of the skin is lined with open-mouthed cells, which pour their gummy secretions into the interior of the carpel, filling it and bathing the seed. The pulp is sweet and aromatic, and is esteemed for making conserves, and also as a cooling drink. In India, the roots and leaves are used medicinally. Bael gum is a sticky, astringent substance soluble in water. The fruit contains several large, flat, woolly seeds, which germinate readily, and the plant is also very easily propagated from root cuttings. Plate LXXXII
- 71. Plate LXXXII.—Bhel or Bael Fruit. One third natural size.
- 72. G. P. W. Collection. Diospyros decandra. BROWN PERSIMMON. This is an evergreen tree rarely found in Hawaii. It has alternate, irregular, long, narrow leaves, shiny dark-green on the upper side, a velvety light-green on the underside, and has a long petiole. The branches are brittle, light-green, smooth and shiny when young, and after the leaves shed become woody and inclined to dry back. The trunk and bark of the tree is covered with warty excresences. The solitary flowers are four-petaled. The edible fruit ripens in December, is round, depressed, about 2½ inches in diameter, in color light-green dotted with numerous white spots. When quite ripe the thin skin turns to a shiny-brown. The soft chocolate colored pulp is sweet and contains from 1 to 8 large flat seeds. Plate LXXXIII
- 73. Plate LXXXIII.—Brown Persimmon. Natural size.
- 74. G. P. W. Collection. Lucuma Rivicoa. EGG FRUIT. This small evergreen tree, which is a native of Brazil, is found only in one or two gardens in Honolulu. Its leaves are elliptic-obovate, resembling those of the mango. The yellow flowers are single, the fruit is the size and shape of a hen's egg, and has the flavor of the yolk of an egg sweetened with sugar. It has from one to three large seeds, which are easily germinated. Plate LXXXIV
- 75. Plate LXXXIV.—Egg Fruit. One third natural size.
- 76. G. P. W. Collection. Eriobotrya Japonica. LOQUAT. The Loquat has been for many years a familiar fruit in our gardens, and is a native of China and Japan. It is a low evergreen tree with thick foliage, and in congenial climates is a profuse bearer. Its leaves are thick, oblong, and remotely toothed and grow near the ends of the branches. The white flowers grow in clusters, are very fragrant, and the fruit, which also ripens in clusters, about Christmas time, is pear-shaped, and has an agreeable acid flavor. The seeds are large, and germinate readily. Fine grafted and budded varieties have been introduced by local horticulturalists. Plate LXXXV
- 77. Plate LXXXV.—Loquat. One fourth natural size.
- 78. G. P. W. Collection. Litchi Chinensis. LICHEE. This tree, with its dense foliage, is a native of Southern China. The first tree of this variety was brought to Hawaii by Mr. Afong, and planted at his residence in Nuuanu avenue, Honolulu, in the year 1870. The leaves are alternate, and abruptly pinnate; the oblong leaflets are not quite opposite. Flowers pale green, small and regular, producing bunches of reddish-colored fruits, each about the size of a small walnut. They are covered with a parchment-like skin having many soft spines. The interior consists of a large seed covered with a whitish pulp of a sweetish acid flavor; this pulp when dried in the shell becomes somewhat shriveled, brownish in color, and very sweet. The fruiting season is in July, and as there are but few trees here that bear, high prices are obtained for this rare fruit, which is much prized by the Chinese. Fresh seeds will germinate, but it requires so many years for these seedlings to bear that grafted and budded plants are imported from China. Plate LXXXVI
- 79. Plate LXXXVI.—Lichee. One third natural size.
- 80. G. P. W. Collection. Euphoria Longana. LONGAN. This tree is a native of India and Southern China. It produces its flowers and fruits at about the same time of year as does the Litchi, which it somewhat resembles, although its fruits are somewhat smaller and less palatable. The tree grows to a height of about 20 feet. It has large, alternate, pinnate leaves, and the oblong leaflets are not quite opposite; they are glossy on the upper surface, and a dusty-brown on the underside. The small flowers come in terminal panicles; and the fruit, which is borne in clusters, has a thin, brittle, somewhat rough shell. There is one large, smooth, hard seed; around which is a thin layer of sweetish, aromatic pulp. The best fruits raised here are those grown by the Chinese. Plate LXXXVII
- 81. Plate LXXXVII.—Longan. One third natural size.
- 82. G. P. W. Collection. Morus nigra. MULBERRY. This low-growing tree is a native of southwestern Russia and Persia. It has rough, dark-green leaves, usually not lobed. The thick, fleshy fruit is variable in size. The mulberry grows readily from cuttings. Plate LXXXVIII
- 83. Plate LXXXVIII.—Mulberry. One third natural size.
- 84. G. P. W. Collection. Garcinia mangostana. MANGOSTEEN. This tree is a native of Sumatra and of the Islands of the Eastern Archipelago. It is of medium size, the stem rising to a height of about 20 feet; and its branches coming out in regular order give the head of the tree the form of a parabola. The leaves are about 8 inches long and 4 inches broad at the middle; they are a beautiful green on the upper side and a delicate olive on the under side. The flowers resemble a single rose with dark-red petals. The fruit is round, about the size of a small orange, and has a characteristic persistent calyx. The shell is at first green, and when ripe changes to purplish-brown marked with yellow spots. The Mangosteen is called the queen of fruits, and the tree upon which it is produced is most graceful and beautiful. Those who have tasted this fruit in its perfection declare it to be indescribably delicious. The Mangosteen must have a hot, moist, and fairly equable climate throughout the year. Many Mangosteen trees have been brought to Hawaii, and have received intelligent care, but they have not thrived well; and have eventually died. Only two have ever produced fruit; one in the garden of Mr. Francis Gay of Kauai, which bears its fruit annually, and the other tree at Lahaina, Maui, in the garden formerly the property of Mr. Harry Turton. Plate LXXXIX
- 85. Plate LXXXIX.—Mangosteen. Two thirds natural size.
- 86. Welcome to our website – the perfect destination for book lovers and knowledge seekers. We believe that every book holds a new world, offering opportunities for learning, discovery, and personal growth. That’s why we are dedicated to bringing you a diverse collection of books, ranging from classic literature and specialized publications to self-development guides and children's books. More than just a book-buying platform, we strive to be a bridge connecting you with timeless cultural and intellectual values. With an elegant, user-friendly interface and a smart search system, you can quickly find the books that best suit your interests. Additionally, our special promotions and home delivery services help you save time and fully enjoy the joy of reading. Join us on a journey of knowledge exploration, passion nurturing, and personal growth every day! ebookbell.com