![Figures
1.1 Typical schematic sagittal representation of the “speech”
apparatus page 20
1.2 Observable acoustic structures of an utterance 25
2.1 Examples of the speech stimuli used by Boucher et al. (2018) in
a sound–picture association task 48
2.2 Learners’ sound–picture associations across trials and
feedback conditions (Boucher et al., 2018) 48
3.1 Examples of formal syntactic analyses performed on
orthographic units 54
4.1 Tracings of radiographic recordings for [ku] and [ki] 78
4.2 Tracings of radiographic recordings for a spontaneous speech
error 83
4.3 Tracings of radiographic recordings for a corrected form
following a spontaneous speech error 83
4.4 Faber’s (1990) list of early Greek letters with their precursors
in Old Aramaic script 86
4.5 Basic spectrographic values that are sufficient to synthesize /di/
and /du/, from Liberman et al. (1967) 90
4.6 Katakana signs representing how similar-sounding syllables can
be variably interpreted depending on the choice of writing
systems 96
4.7 Illustration of dynamic characteristics of speech production 102
6.1 A general scenario where a speaker offers food to a listener
while producing utterances 119
6.2 The relationship between acoustic stimuli consisting of
a synthesized “da” and ABRs (according to Skoe & Kraus, 2013) 130
6.3 Functional effects of the phase of neural oscillations
(according to Schroeder et al., 2008) 136
7.1 Representations and transformations from input signal to
lexical representation as viewed by Poeppel et al. (2007) 145
7.2 An illustration of the EP hypothesis of intrinsic force-length
relationships in muscle control 148
ix](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-15-320.jpg)
![7.3 Idealized unitary mass displacement in a critically damped and
under-damped spring system at three times (Boucher, 2008) 150
7.4 Guenther’s DIVA control scheme (2016) 153
7.5 Rectified and smoothed EMG activity of opener and closer
muscles of lip and jaw motions, along with midline articulator
displacements, bilabial compression, and intra-oral pressure
(Boucher, 2008) 159
7.6 Velocity and mid-sagittal displacement of lip and jaw opening
as a function of force-related measures of closing motions
(Boucher, 2008) 160
7.7 Fiber composition of intrinsic tongue muscles in four parts of
the tongue (Stål et al., 2003) 161
7.8 Glottal motion and EMG activity of laryngeal muscles
preceding the first vowel in “I say” (Hirose et al., 1980) 163
7.9 Overall correct serial recall of auditorily presented CVand VC
sequences of nonsense syllables 165
7.10 Audio signal, intra-oral pressure, and kinematics of midline
labial motions during the production of Bobby [‘bɑbi] and
poppy [‘pɑpi] 167
8.1 VOTs for [ta] and [da] produced at varying rates of speech
(Boucher, 2002) 174
8.2 Pitch (F0), and dB energy patterns of utterance stimuli along
with corresponding ERPs of regions of interests (Gilbert et al.,
2015) 178
8.3 Inter-trial phase coherence (ITPC) for three types of stimuli as
a function of the frequency of oscillations (Boucher et al.,
2019) 180
9.1 Age-related changes in speech breathing recorded via
plethysmographic belts 186
9.2 Vital capacities in liters and MLU in morphemes and syllables
(Boucher & Lalonde, 2015) 189
9.3 MLU in morphemes and syllables as a function of vital
capacity (Boucher & Lalonde, 2015) 189
9.4 Overall percentages of nominal forms of 1, 2, and 3 syllables or
more used by 50 speakers aged 5–27 years (Boucher &
Lalonde, 2015) 191
10.1 An illustration of the embodied approach in which the
semantics of the lexeme banana is seen to be grounded in
perception and action systems (Kemmerer, 2015) 210
10.2 Effects of rehearsing lexical items in different production
conditions on participants’ recall of having produced the items
(Lafleur & Boucher, 2015) 216
x List of Figures](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-16-320.jpg)
![Geneva lectures of 1907–11” (p. 21). The speech–language division is a prevailing
doctrine, and also reflects nineteenth-century views of a hierarchical brain
(Parvizi, 2009). However, the original claim of a physiological division no longer
holds, and arguments for the separate study of language have been criticized on
fundamental grounds.
1.3.1 Saussure’s Argument of a Separate Language Faculty
in Broca’s Area
In terms of neurophysiological evidence, at the time of his Cours de linguis-
tique générale, Saussure believed that the capacity to create a system of signs
was located in Broca’s area (Broca, 1861). In discussing this function, Saussure
distinguished language from the motor-control processes of parole, though he
did not attempt to localize the latter. In citing Saussure’s Cours, it is worth
mentioning that the English translation by W. Baskin alters key terms in that
“speech” is often used where Saussure uses langage, so the following excerpt
restitutes the term “language” (in brackets) that is in the French version of the
Cours. It is perhaps useful to note that compared to English, which has two
terms to designate language and speech, French has four: parole, langue,
langage, and un parler. In Saussure’s view, language (langage) is the capacity
to constitute a system of coded signs that can be a spoken code or langue,
a writing code, or other systems of signs including gestural “sign languages.”
Saussure’s central claim is that the division between speech and language is not
merely methodological, but physiological:
We can say that what is natural to mankind is not oral speech but the faculty of
constructing a language, i.e. a system of distinct signs corresponding to distinct ideas.
[. . .]
Broca discovered that the faculty of [language] is localized in the third left frontal
convolution; his discovery has been used to substantiate the attribution of a natural
quality to [language]. But we know that the same part of the brain is the center of
everything that has to do with [language], including writing. The preceding statements,
together with observations that have been made in different cases of aphasia resulting
from lesion of the centers of localization, seem to indicate: (1) that the various disorders
of oral [language] are bound up in a hundred ways with those of written [language]; and
(2) that what is lost in all cases of aphasia or agraphia is less the faculty of producing
a given sound or writing a given sign than the ability to evoke by means of an
instrument, regardless of what it is, the signs of a regular system of [language]. The
obvious implication is that beyond the functioning of the various organs there exists
a more general faculty which governs signs and which would be the linguistic faculty
proper. (Saussure, 1916/1966, pp. 10–11)
The idea of a separate language faculty is still present in the literature. Yet
setting aside the historical debate on localizing aphasia (see, e.g., Ross, 2010),
it has been recognized that the notion of a circumscribable language function in
21
How We Are Introduced to the Study of Spoken Language](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-45-320.jpg)
![that spoken language involves an autonomous system located in Broca’s area,
and saw this system as physiologically separate from speech. Based on these
assumptions, he proposed methods by which to analyze the abstract system
using transcripts as he saw no need to refer to the structural aspects of
motor-sensory processes of vocal expression. To the contrary, research
shows that there is no cortically localized language function, and that
“language” and motor-sensory processes intertwine. There are several impli-
cations for Saussure’s approach to the study of spoken language via tran-
scripts, one of which should be clear: if there is no physiological basis for
dividing language from speech, then observable structures of utterances are
the structures of spoken language. In fact, the natural structures of spoken
language were never “given” or represented by culture-specific writing signs
and units. In continuing to idealize language via transcripts and concepts
relating to letters, words, sentences (etc.), analysts fail to recognize that the
principal historical justification for using script in the study of spoken lan-
guage was the nineteenth-century view of an autonomous language faculty,
which no longer holds.
1.3.2 Arguments of the Arbitrariness of Signs and Abstract Phonology
Saussure presented several influential arguments supporting a Cartesian view
of language that still prevails. In particular, Saussure taught in his Cours that
the capacity to create arbitrary (symbolic) sound–meaning associations and
phonological systems demonstrates that a language function is separate from an
oral modality of communication and is solely governed by what he called
“rational” principles (which can be seen to refer to cognitive functions before
G. Miller formally introduced the term).
The linguistic sign is arbitrary; language, as defined, would therefore seem to be
a system which, because it depends solely on a rational principle, is free and can be
organized at will. (p. 78)
[. . .]
Just what phonational movements could accomplish if language did not exist is not
clear; but they do not constitute language, and even after we have explained all the
movements of the vocal apparatus necessary for the production of each auditory impres-
sion, we have in no way illuminated the problem of language. It is a system based on the
mental opposition of auditory impressions, just as a tapestry is a work of art produced by
the visual oppositions of threads of different colors; the important thing in analysis is the
role of the oppositions, not the process through which the colors were obtained. (p. 33)
The quoted assertions are characteristic of the Cours. It is the case that signs in
spoken language appear arbitrary and that distinctive features vary across lan-
guages. But the conclusion does not follow that signs and feature systems are
23
How We Are Introduced to the Study of Spoken Language](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-47-320.jpg)
![a firewall against invalidating evidence, as Jackendoff (2009) has remarked.
One critical point of disconnect between theoretical concepts of language
analysis and empirical observations occurred with the rise of phonology.
1.3.3 On the Primacy of Linguistic Criteria: The Historical Disconnect
from Instrumental Observations
In considering language analysis, one might wonder why readily observed
structures, as visualized in Figure 1.2 and which are universal in speech, are
not generally viewed as “design features” of spoken language (Hockett, 1960).
A principal historical reason for this oversight may be that the structural
patterns are not easily transcribed. On this point, it is essential to note that,
when methods of linguistic analysis were being devised, the basic instruments
used today to record and view speech signals were not widely available.
Saussure, for instance, was a philologist who worked on ancient texts and
lectured on the reconstruction of proto-Indo-European. At the time, transcripts
were the main data if not the only data on which to formulate an understanding
of spoken language. Instruments that could visually record speech, such as the
kymograph (Vierordt & Ludwig, 1855), were rarely used, and the first spectro-
graph only appeared decades later (Steinberg, 1934). But there was also some
reticence as to the usefulness of instrumental observations, even amongst
classical phoneticians. The reason was that both philologists and phoneticians
were applying subjective classifications of speech sounds and many did not see
the need for instruments (e.g., in 1913, Jespersen saw that “Les instruments qui
permettent de mesurer les différences de position des organes et de représenter
par des chiffres la tension des muscles ou l’amplitude des ondes [. . .] sont
inutiles”; in Auroux, 2000, p. 511; and see Malmberg, 1972). A similar disin-
terest for instrumental methods prevailed amongst early phonologists of the
Figure 1.2 Observable acoustic structures of an utterance: Cyclical patterns
in energy contours, temporal groupings, tonal contours, and both amplitude
and pitch declination across breath units of speech.
25
How We Are Introduced to the Study of Spoken Language](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-49-320.jpg)
![Prague School in the 1920s, which included such influential authors as
Trubetzkoy (1890–1938), Jakobson (1896–1982), and adherents like Martinet
(1908–1999), Benveniste (1902–1976), among others (see Fischer-Jørgensen,
1975). How, then, did analysts come to conceptualize systems of sound oppos-
ition as put forward by Saussure without sound-recording instruments?
A reading of early works offers a view of how the analyses were performed. At
the time of Saussure’s Cours, philologists were engaged in comparing languages
often using ad hoc alphabet-style symbols. With a view to reconstructing proto-
Indo-European, analysts of ancient texts focused on changes in sound distinc-
tions, which were described using impressionistic articulatory categories. Terms
like “stop,” “fricative,” “front,” “back,” “palatal” (etc.) were broadly used by
both philologists and phoneticians like Passy (1859–1940) and Sweet (1845–
1912). But it is worth noting that these categories were subjectively established.
Thus, when classical phoneticians categorized sounds according to articulatory
place and mode, they did so by reference to what they felt they were doing when
producing the sounds. Moreover, in an effort to standardize descriptions, Sweet
(1877) suggested that different signs should be used only for “those broader
distinctions of sound which actually correspond to distinctions of meaning in
a given language” (p. 103), although this was also problematic. To illustrate the
way speech was being described, consider how Sweet (1877) discussed the
difficulty in segmenting sounds (in this case “glides”):
Consonant-glides are more noticeable in French than in English, especially in stop-
combinations, (strik[ʜ]t)=‘strict.’ Final voice stops often end in a voice-glide, (bag[ʌ])
=‘bague.’ In passing from (ɴ) to the next vowel the glide is generally formed so slowly
as to be heard as a separate element, so that (ohɴ[i]oq)=‘oignon’ sounds like (ohɴjoq).
Final (j) and (ɴ) end voicelessly, the glottis being opened at the moment of removing the
tongue from the consonant position, so that (fijʜ) and (vIɴʜ) sound like (fij-jh) and (viʜ-
jh). (p. 125)
In later attempts to standardize ad hoc signs, alphabet systems came to be used.
However the choice of a letter system was circumstantial. In 1886, Passy headed
a group of phoneticians and founded the Association phonétique internationale
with the purpose of developing a transcription system for teachers of European
languages (Auroux, 2000). In 1888, the association published the first version of
the International Phonetic Alphabet, which came to be adopted by philologists
and phonologists. The latter formally introduced a “commutation test” as a method
for establishing the phonological system of a language. With this method, IPA
letters are substituted in “minimal pairs” to identify “distinctive features”
that differentiate literal meanings in paired lexemes. For instance, in English
/bak/-/pak/, the commutation of b-p is seen to create a distinction in meaning in
terms of a subjectively estimated “voice” feature. However, letters in such analyses
were not merely a descriptive device. They took on a theoretical and psychological
26 Questions of Ontology](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-50-320.jpg)
![V does not start after a consonant C in producing CV). His report also showed
that one could not segment speech into units smaller than syllable-like
“ballistic pulses” – that is, one cannot halt speech during a pressure-
building constriction of the vocal tract or the vocal folds such that speech
minimally involves a cycle of motions that build and release pressure. Stetson
explicitly outlined the implications for language analysts by quoting from
works in phonology:
To isolate and classify the essential “sounds” (phonemes) of a language, i.e. to assemble
the phonetic alphabet of a language, was an important project for Sweet and F. de
Saussure, and a primary enterprise for Trubetzkoy. This was the impulse that lay behind
the International Phonetic Alphabet, the first achievement of the phoneme doctrine.
Sweet noted that “sounds” differentiated words; at the hands of Saussure, Trubetzkoy
and associates, the smallest phonic change in a word that shifts the meaning was made to
indicate a new phoneme.
“Dans le ‘projet de terminologie standardisée’ soumis à la Réunion Phonologique
Internationale de 1930, on trouve les définitions suivantes: ‘Une opposition pho-
nologique est une différence phonique susceptible de servir dans une langue
donnée à la différenciation des significations intellectuelles; chaque terme d’une
opposition phonologique quelconque est une unité phonologique; le phonème est
une unité phonologique non susceptible d’être dissociée en unités phonologiques
plus petites et plus simples.’ (Travaux du Cercle Linguistique de Prague, IV,
p. 311).”
It is apparent that the method involves a resort to meanings and also to observations of
articulations. Certainly, in one form or another, differentiating the significant articula-
tions by differences of meaning will be important to any system. It is unfortunate that the
scholars who have made the most of the method have not only ignored the syllable but
have also insisted that the articulation occurs in the separate, concrete world of la parole,
while the phoneme symbol occurs in the separate, ideal world of la langue; and so the
phonemicists have made a virtue of their ignorance of experimental methods of obser-
vation. (p. 136)
More importantly, Stetson cautioned language theorists on the writing bias
underlying a conceptualization of speech as sequences of letter-like phonemes,
and indicated why these units may be impossible in terms of temporal con-
straints on speech processing.
The series of characters which we read and write as representing an articulate language
have given us a mistaken notion of the units which we utter and which we hear. The
series of speech units cannot correspond to the series of “sounds” or “phonemes” set
down on paper. Even “slow, careful utterance,” let alone the rapid utterance of everyday,
is much too fast for that. The maximum rate at which articulations can be uttered is
10–12 per sec.; and the maximum rate at which auditory signals can be identified is
14–16 per sec. A [transcribed] syllable of speech often indicates 2–7 phonemes; “do”
has two, “tree” three, “quilt” five, “squelched” seven; the slow, careful rate of utterance
is often 4–5 syllables per sec. Thus phonemes are often indicated 15–25 per sec.
Obviously the phonemes are not uttered (or heard) one after the other; there must be
28 Questions of Ontology](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-52-320.jpg)
![extensive overlapping, as physiological tracings prove. The consonants do not prove to
be separable units, they must have breath pressure behind them; the pressure is supplied
by the pulses of the syllables in which they function. A consonant cannot be pronounced
alone; it is always a characterized factor in some syllable. The supposed consonant
“elements”, which naive teachers assume that they are uttering, prove to be syllables
with pulses from the chest, but often with the vowel shape unvocalized. (p. 137)
The idea of syllable-like pulses or cycles is developed later on in this mono-
graph, but it is useful to note that Stetson’s wording of “syllables with pulses from
the chest” suggested that the pulses link to respiratory mechanisms. Studies using
electromyography (EMG) have shown, to the contrary, that expiratory processes
are largely passive during normal speech (Draper, Ladefoged, & Whitteridge,
1959; Ladefoged, 1962). According to Ohala (1990, p. 30), Stetson likely misin-
terpreted observations of pressure variations in the lungs as indicating an active
factor when, in fact, the variations reflect a resistance to expiratory flow created by
cyclical articulatory motions upstream (see also Boucher & Lamontagne, 2001).
On the other hand, several studies support Stetson’s main claim: speech cannot
be segmented in units smaller than syllable-like pulses seen as coherent cycles of
pressure-building and releasing motions where the latter can be voiced or
unvoiced. Research in the 60s confirmed that C and V motions are not timed as
successive phonemes, but “co-articulated” in coherent CV-like units (MacNeilage
& DeClerk, 1967, and see Chapter 4). Furthermore, studies combining EMG and
photoglottography confirm cyclical motions and pulses as basic units of produc-
tion. Thus, sounds that can be produced in isolation, like vowels and “syllabic”
consonants, minimally involve a pressure-rising constriction of the vocal folds
prior to a voiced release (Hirose, Sawashima, & Yoshioka, 1980, 1983). Stop
consonants like [p, t, k, b, d, g], however, cannot be produced in isolation.
Articulation of these sounds involves unitary cycles where a pressure-building
oral closure entails a releasing motion, and this is observed in CV, post-vocalic
contexts as in VC, CVC, and complex clusters where the oral release can be
voiceless, or voiced as brief schwas (e.g., Ridouane, 2008). The latter findings,
incidentally, suggest that IPA notations and orthographic notions of VC and CVC
as single syllables misrepresent the fact that these sequences are produced as two
syllable-like cycles of motion each involving separate bursts of EMG activity
(Boucher, 2008, and see the demonstrations in Part III). Stetson’s view of unitary
cyclical pulses can account for several phenomena such as why certain consonant
motions can create syllable beats, or the particular coherence of CV patterns
(MacNeilage & DeClerk, 1967), discussed in subsequent chapters. As for the
remark on the impossibility of phoneme-by-phoneme processing in production
and perception, Stetson’s claim of a maximal rate of about 100 ms/syllable
conforms to an observed upper limit on the production of separate sounds (Kent,
1994; Knuijt, Kalf, Van Engelen et al., 2017). In terms of acoustic perception, an
extensive review by Warren (2008) indicates that order perception for unstructured
29
How We Are Introduced to the Study of Spoken Language](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-53-320.jpg)
![speech sounds such as sequences of vowels or digits separated by silences is
possible for rates no lower than 100 ms per item. Sequences in which items are
artificially compressed to about phoneme-size elements can create auditory illu-
sions and confusions in perceived order. If such observations are confirmed for
structured speech, they would further substantiate Stetson’s original critique of
phonemes as processing units.
But findings such as those reported by Stetson were not a serious concern for
language theorists of the day. In overlooking instrumental findings, a favored
argument of early phonologists that is often repeated in introductory texts, was
that instrumentalists could describe speech sounds to no end. Consequently,
some saw instrumental observations as having to be guided by considerations
of the functional aspects of sounds serving to communicate meaning distinctions.
This argument of the “primacy of linguistic criteria” was present in Saussure’s
writings, in early phonetics (see Sweet’s notion of “broad” transcriptions), and
was essential to the constitution of phonology as a separate subfield of language
study. As Trubetzkoy (1939/1971) stated: “It is the task of phonology to study
which differences in sound are related to differences in meaning in a given
language, in which way the discriminative elements [. . .] are related to each
other, and the rules according to which they may be combined into words and
sentences” (p. 10). For some phonologists this meant that speech observations
had to be hypothesis driven, they had to follow assumptions of linguistic analyses
and the criteria of distinctiveness because “nothing in the physical event . . . tells
us what is worth measuring and what is not” (Anderson, 1985, p. 41, quoted by
Lindblom, 2000; see also Anderson, 1981).
It is the case that, to communicate using spoken language, sounds need to be
distinctive. But as noted, nothing in the principle of distinctiveness provides an
epistemological justification for restricting “features” to sounds that can be
transcribed, or to indices that create differences in literal meaning, such as in
analyses of minimal pairs of “words” as listed in dictionaries. Nor does this
principle serve to justify the view that distinctive features occur in letter-like
segments, or any other extent of speech for that matter. The issue of sound
segmentation is an empirical question bearing on how features are produced
and perceived along a time axis. It is an issue that was not resolved by the
choice of alphabet signs to represent speech. On the contrary, in the history of
language study, the decision to analyze speech as sequences of letters said to
represent phonemes finds no “phonological” or experimental motivation. Yet in
choosing to represent sound features with letters, the notion of letter-like
packets of features gained a theoretical status, even though this choice was
entirely circumstantial. In fact, when the field of language study began to
consolidate in the latter part of the nineteenth century, various transcription
systems were available. Some systems, like Bell’s Visible Speech, which
became popular in some circles, did not involve letters at all (Abercrombie,
30 Questions of Ontology](https://image.slidesharecdn.com/11974137-250528091450-fbbd8690/85/The-Study-Of-Speech-Processes-Addressing-The-Writing-Bias-In-Language-Science-Victor-J-Boucher-54-320.jpg)
The Study Of Speech Processes Addressing The Writing Bias In Language Science Victor J Boucher
- 1. The Study Of Speech Processes Addressing The Writing Bias In Language Science Victor J Boucher download https://ebookbell.com/product/the-study-of-speech-processes- addressing-the-writing-bias-in-language-science-victor-j- boucher-23948274 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. Languagean Introduction To The Study Of Speech 2001th Edition Edward Sapir https://ebookbell.com/product/languagean-introduction-to-the-study-of- speech-2001th-edition-edward-sapir-59346460 The Logic Of Language A Semiotic Study Of Speech Michael Shapiro https://ebookbell.com/product/the-logic-of-language-a-semiotic-study- of-speech-michael-shapiro-44938906 Speech Play Research And Resources For The Study Of Linguistic Creativity Reprint 2016 Barbara Kirshenblattgimblett Editor https://ebookbell.com/product/speech-play-research-and-resources-for- the-study-of-linguistic-creativity-reprint-2016-barbara- kirshenblattgimblett-editor-51963562 Approaches To The Study Of Sound Structure And Speech Interdisciplinary Work In Honour Of Katarzyna Dziubalskakoaczyk Magdalena Wrembel https://ebookbell.com/product/approaches-to-the-study-of-sound- structure-and-speech-interdisciplinary-work-in-honour-of-katarzyna- dziubalskakoaczyk-magdalena-wrembel-33351946
- 3. Approaches To The Study Of Sound Structure And Speech Magdalena Wrembelagnieszka Kiekiewiczjanowiakpiotr Gsiorowski https://ebookbell.com/product/approaches-to-the-study-of-sound- structure-and-speech-magdalena-wrembelagnieszka- kiekiewiczjanowiakpiotr-gsiorowski-232027602 Lingua Indisciplinata A Study Of Transgressive Speech In The Romance Of The Rose And The Divine Comedy Phd Diss Gabriella Ildiko Baika https://ebookbell.com/product/lingua-indisciplinata-a-study-of- transgressive-speech-in-the-romance-of-the-rose-and-the-divine-comedy- phd-diss-gabriella-ildiko-baika-5848976 Illocutionary Constructions In English Cognitive Motivation And Linguistic Realization A Study Of The Syntactic Realizations Of The Directive Commissive And Expressive Speech Acts In English 1st Edition Nuria Del Campo Martnez https://ebookbell.com/product/illocutionary-constructions-in-english- cognitive-motivation-and-linguistic-realization-a-study-of-the- syntactic-realizations-of-the-directive-commissive-and-expressive- speech-acts-in-english-1st-edition-nuria-del-campo-martnez-5595656 The Study Of Words An Introduction Lewis Gebhardt https://ebookbell.com/product/the-study-of-words-an-introduction- lewis-gebhardt-47644074 The Study Of Language 5th Edition 5th Edition George Yule https://ebookbell.com/product/the-study-of-language-5th-edition-5th- edition-george-yule-48038870
- 7. The Study of Speech Processes There has been a longstanding bias in the study of spoken language toward using writing to analyze speech. This approach is problematic in that it assumes language to be derived from an autonomous mental capacity to assemble words into sentences, while failing to acknowledge culture- specific ideas linked to writing. Words and sentences are writing constructs that hardly capture the sound-making actions involved in spoken language. This book brings to light research that has long revealed structures present in all languages but which do not match the writing-induced concepts of trad- itional linguistic analysis. It demonstrates that language processes are not physiologically autonomous, and that speech structures are structures of spoken language. It then illustrates how speech acts can be studied using instrumental records, and how multisensory experiences in semantic memory couple to these acts, offering a biologically grounded understanding of how spoken language conveys meaning and why it develops only in humans. victor j. boucher is Senior Researcher and Professor of Speech Sciences at the Université de Montréal. His career work on the physiological processes of speech have led him to view human language as arising from constraints on motor-sensory systems and to a critical reappraisal of methods of language study.
- 9. The Study of Speech Processes Addressing the Writing Bias in Language Science Victor J. Boucher Université de Montréal
- 10. University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 79 Anson Road, #06–04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781107185036 DOI: 10.1017/9781316882764 © Victor J. Boucher 2021 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2021 A catalogue record for this publication is available from the British Library. ISBN 978-1-107-18503-6 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
- 11. Contents List of Figures page ix List of Tables xii List of Abbreviations xiii Preface xv Introducing a Fundamental Problem of Language Science 1 Part I Questions of Ontology: Writing and the Speech–Language Divide 11 1 How We Are Introduced to the Study of Spoken Language 13 1.1 Language as an “Autonomous” System, or the Effects of Scriptism 14 1.2 Defining “Speech” 18 1.3 Was the Speech–Language Division Ever Physiologically Grounded? 20 1.3.1 Saussure’s Argument of a Separate Language Faculty in Broca’s Area 21 1.3.2 Arguments of the Arbitrariness of Signs and Abstract Phonology 23 1.3.3 On the Primacy of Linguistic Criteria: The Historical Disconnect from Instrumental Observations 25 1.3.4 Explaining Systems of Distinctive Features: Lindblom’s Demonstration (1986) 31 2 The Modality-Independence Argument and Storylines of the Origin of Symbolic Language 34 2.1 Cognitive Skills as Insufficient Factors in the Rise of Symbolic Communication 35 2.2 The Case against Modality-Independent Accounts of Symbolic Language 37 2.3 Modality-Dependent Accounts of the Rise of Symbolic Language 41 2.3.1 Mimesis, Procedural Learning, and the Case of Sign Languages 41 2.3.2 “Sound Symbolism”: Questions of the Efficiency of Iconic Signs 44 2.3.3 Articulated Vocalization and the Rise of Symbolic Signs: A Laboratory Demonstration 45 2.4 The Phylogeny and Ontogeny of an Amodal Symbol Function as a Pseudo-Puzzle 49 v
- 12. 3 The Recent History of Attempts to Ground Orthographic Concepts of Language Theory 52 3.1 From Orthographic Representations to “Substantive Universals” 53 3.2 Shoehorning Orthographic Concepts: Issues in Grounding the LAD 58 3.2.1 Biases and Limitations of Analyzing Language Development through Writing 59 3.2.2 The Search for Marks of Words and Phrases, versus “Chunks” 61 3.3 Neuroscience Falls upon Nonexistent Substantive Universals: Why This Invalidation Is Different 66 3.4 Abandoning the Competence–Performance Divide 69 Postscript – On the Use of the IPA and Terms of Latin Grammar in the Present Work 70 Part II Questions of Epistemology: The Role of Instrumental Observations 73 4 Recognizing the Bias 75 4.1 On the Tradition of Overlooking Instrumental Observations: The Case of the Phoneme 76 4.1.1 From Instrumental Records of Co-articulation to Transcribed Spoonerisms 77 4.1.2 On the Origin of Alphabet Signs: The Hypothesis of a Preliterate Awareness of Phonemes 85 4.1.3 Testing Phoneme Awareness: Issues in Defining Reference Units 89 4.2. The Looking-Glass Effect: Viewing Phoneme Awareness by Reference to IPA Transcripts 94 4.2.1 “Phonological” Evidence of Phonemes Versus Motor Processes 99 4.2.2 On Arguments of the “Logical Necessity” of Phonemes and the Success of Alphabet Systems 103 4.2.3 Effects of Writing on Speakers’ Awareness of Words, Phrases, Sentences 105 5 (Re-)defining the Writing Bias, and the Essential Role of Instrumental Invalidation 108 5.1 On the Persistence of Scriptism in the Study of Spoken Language 108 5.2 The Need to Address Complaints of Cultural Centrism and Ethical Concerns 111 Part III The Structure of Speech Acts 115 6 Utterances as Communicative Acts 117 6.1 Describing Speech Acts and Their Meaning 117 6.2 The Parity Condition, Motor-Sensory Coupling, and the Issue of Utterance Structure 125 6.3 The Coding of Speech Acoustics in the Auditory Brain Stem and Effects of Motor-Sensory Coupling 129 6.4 Multimodal Sensory Integration: Introducing Neural Entrainment to Speech Structure 133 vi Contents
- 13. 6.4.1 The Specificity of Neural Entrainment in the Speech Modality 133 6.4.2 Neural Entrainment to Structures of Motor Speech: Linking to Spiking Activity 134 6.4.3 On the Role of Subcortical Processes: Multisensory-to-Motor Integration and Chunking 137 6.5 Relating to Utterance Structure, or What the Brain Does Not Intrinsically Construct 141 7 Relating to Basic Units: Syllable-Like Cycles 143 7.1 Speech Production: On the Brain–Utterance Interface That Never Was 143 7.2 Basic Sequencing Units in Theories of Speech-Motor Control: Some Examples 147 7.2.1 The Equilibrium-Point (EP) Hypothesis 147 7.2.2 The Task Dynamics (TD) Model 149 7.2.3 Directions in Auditory Space Into Velocities of Articulators: The DIVA Model 152 7.3 Critical Evidence of Basic Sequencing Units and What Shapes Them 156 7.3.1 Intrinsic Muscle-Tissue Elasticity and Its Effect on Speech Motions 157 7.3.2 Other Intrinsic Effects of Muscle Tissues on Motion Sequencing within Syllable Cycles 159 7.3.3 Just How Many Units Are There in CV and VC, and Are These Represented in Memory? 162 7.3.4 Syllable Cycles within Chunks and Graded Motion Control without Phonemes 165 8 Relating Neural Oscillations to Syllable Cycles and Chunks 172 8.1 The Entrainment of Low-Frequency Oscillations and Speech Processing 172 8.1.1 On the Role of Theta- and Delta-Size Processing Windows 173 8.1.2 Reviewing Claims of a Non-sensory Entrainment of Delta to Content Units 174 8.2 Delta-Size Windows and the Sensory Chunking of Speech 176 8.2.1 Chunks and Their Signature Marks 176 8.2.2 Neural Entrainment in Speech Processing 179 9 Breath Units of Speech and Their Structural Effects 182 9.1 Utterances as Breath Units versus Sentences in Speaker–Listener Interaction 182 9.2 On Interpreting Measures of “Mean Length of Utterance” (MLU) 183 9.2.1 Utterance Complexity, Lexical Diversity, and MLU: Linking to Developing Motor Structures 187 9.2.2 Chunks in Breath Units of Speech and the Development of Vocabulary 190 9.2.3 On Explaining Developmental Milestones 192 9.3 The Structure of Spoken Language: An Interim Summary with a View on Addressing the Issue of Scriptism 194 Part IV The Processing of Speech Meaning 197 10 The Neural Coding of Semantics 199 10.1 Units of Writing, Structures of Utterances, and the Semantics of Speech 199 vii Contents
- 14. 10.2 The Lexico-Semantic Approach: Context Information as “Nonessential” 200 10.2.1 Lexico-Semantics and Traditional Models of Language Processing 201 10.2.2 Embodied versus Disembodied Semantics 203 10.3 How Semantic Representations of Verbal Expressions Develop: On “Modes of Acquisition” 205 10.4 The Partitioning of Semantic Memory and Its Formatting in Spoken Languages 208 10.4.1 Words Are Not Biologically Grounded Units: Why Sensory Chunking Is Necessary 211 10.4.2 On Representations of Verbalized Forms in Memory: Activating Episodes of Speech Acts 214 10.5 The Nature of Semantic Representations: On the Neural Coding of Context Information in Action Blocks of Speech 216 11 Processes of Utterance Interpretation: For a Neuropragmatics 220 11. 1 The Issue of the Selective Activation of Semantic Representations in Speech Contexts 220 11.1.1 Context-Based Semantics: Clinical Observations Using Unconventional Test Batteries 222 11.2 On Context-Based Speech Comprehension: Selective Activation of Semantic Representations On-Line 225 11.2.1 Thalamocortical Interactions and the Integrating Role of the Motor Thalamus 227 11.2.2 The Semantics of Utterances: The Analogy of Action Selection in Spatial Navigation 230 11.2.3 Subcortical Mechanisms of Buffering and Context-Based Semantic Processing 234 Epilogue 239 References 244 Index 305 viii Contents
- 15. Figures 1.1 Typical schematic sagittal representation of the “speech” apparatus page 20 1.2 Observable acoustic structures of an utterance 25 2.1 Examples of the speech stimuli used by Boucher et al. (2018) in a sound–picture association task 48 2.2 Learners’ sound–picture associations across trials and feedback conditions (Boucher et al., 2018) 48 3.1 Examples of formal syntactic analyses performed on orthographic units 54 4.1 Tracings of radiographic recordings for [ku] and [ki] 78 4.2 Tracings of radiographic recordings for a spontaneous speech error 83 4.3 Tracings of radiographic recordings for a corrected form following a spontaneous speech error 83 4.4 Faber’s (1990) list of early Greek letters with their precursors in Old Aramaic script 86 4.5 Basic spectrographic values that are sufficient to synthesize /di/ and /du/, from Liberman et al. (1967) 90 4.6 Katakana signs representing how similar-sounding syllables can be variably interpreted depending on the choice of writing systems 96 4.7 Illustration of dynamic characteristics of speech production 102 6.1 A general scenario where a speaker offers food to a listener while producing utterances 119 6.2 The relationship between acoustic stimuli consisting of a synthesized “da” and ABRs (according to Skoe & Kraus, 2013) 130 6.3 Functional effects of the phase of neural oscillations (according to Schroeder et al., 2008) 136 7.1 Representations and transformations from input signal to lexical representation as viewed by Poeppel et al. (2007) 145 7.2 An illustration of the EP hypothesis of intrinsic force-length relationships in muscle control 148 ix
- 16. 7.3 Idealized unitary mass displacement in a critically damped and under-damped spring system at three times (Boucher, 2008) 150 7.4 Guenther’s DIVA control scheme (2016) 153 7.5 Rectified and smoothed EMG activity of opener and closer muscles of lip and jaw motions, along with midline articulator displacements, bilabial compression, and intra-oral pressure (Boucher, 2008) 159 7.6 Velocity and mid-sagittal displacement of lip and jaw opening as a function of force-related measures of closing motions (Boucher, 2008) 160 7.7 Fiber composition of intrinsic tongue muscles in four parts of the tongue (Stål et al., 2003) 161 7.8 Glottal motion and EMG activity of laryngeal muscles preceding the first vowel in “I say” (Hirose et al., 1980) 163 7.9 Overall correct serial recall of auditorily presented CVand VC sequences of nonsense syllables 165 7.10 Audio signal, intra-oral pressure, and kinematics of midline labial motions during the production of Bobby [‘bɑbi] and poppy [‘pɑpi] 167 8.1 VOTs for [ta] and [da] produced at varying rates of speech (Boucher, 2002) 174 8.2 Pitch (F0), and dB energy patterns of utterance stimuli along with corresponding ERPs of regions of interests (Gilbert et al., 2015) 178 8.3 Inter-trial phase coherence (ITPC) for three types of stimuli as a function of the frequency of oscillations (Boucher et al., 2019) 180 9.1 Age-related changes in speech breathing recorded via plethysmographic belts 186 9.2 Vital capacities in liters and MLU in morphemes and syllables (Boucher & Lalonde, 2015) 189 9.3 MLU in morphemes and syllables as a function of vital capacity (Boucher & Lalonde, 2015) 189 9.4 Overall percentages of nominal forms of 1, 2, and 3 syllables or more used by 50 speakers aged 5–27 years (Boucher & Lalonde, 2015) 191 10.1 An illustration of the embodied approach in which the semantics of the lexeme banana is seen to be grounded in perception and action systems (Kemmerer, 2015) 210 10.2 Effects of rehearsing lexical items in different production conditions on participants’ recall of having produced the items (Lafleur & Boucher, 2015) 216 x List of Figures
- 17. 10.3 Phase-amplitude coupling between theta-band oscillations and gamma oscillations (Canolty et al., 2006) 218 11.1 Analyses of LFPs in the subthalamic nucleus by Watson and Montgommery (2006) 236 E.1 An illustration of the relationship between neural responses, assumed “interim” units of language analysis, and speech stimuli 242 xi List of Figures
- 18. Tables 1.1 Part of Lindblom’s (1986) predictions of “vowel systems” using a maximal perceptual distance metric and two acoustic-to -auditory mapping approaches page 32 11.1 Test batteries suggested by Murdoch and Whelan (2009) to observe the effects of pallidotomy, thalamotomy, and pathologies of the cerebellum on “high level linguistics” 224 xii
- 19. Abbreviations ABR auditory brainstem response ABSL Al-Sayyid Bedouin Sign Language ASL American Sign Language CNS central nervous system EEG, MEG electroencephalography, magnetoencephalography EGG electroglottography EMG electromyography EMMA electromagnetic articulography ERP event-related potential F0 fundamental frequency perceived as pitch F1, F2 formants FAF frequency altered feedback FFR frequency following response IPA International Phonetic Alphabet ITPC inter-trial phase coherence LAD Language Acquisition Device LCA last common ancestor LFP local field potential MEP motor-evoked potential MLU mean length of utterance MRI, fMRI magnetic resonance imaging, functional magnetic resonance imaging tDCS transcranial direct current stimulation TMS transcranial magnetic stimulation TP transition probability between transcribed units VOT voice onset time VC vital capacity xiii
- 21. Preface It is customary in a preface to indicate the source of a book, why it was written and for whom, and to thank those who helped forwarding the work through to publication. The present monograph is primarily directed at students and researchers in sectors relating to spoken language. It is also aimed at any interested reader wishing to understand the historical course and recent devel- opments of language science extending to techniques of neuroscience. The book addresses a long-standing problem that may be apparent to anyone with minimal training in methods of linguistic analysis. Such training is part of introductory courses that are often a prerequisite in subprograms of psych- ology, language neuroscience, communication disorders, and language teach- ing, among other disciplines. The tradition has been that all who engage in the study of language are trained in analyzing transcribed speech and thus come to conceptualize spoken language by reference to theories erected on these ana- lyses. This has definite consequences across sectors. By such training, many view spoken language as containing letter- and word-like units, organized in terms of given categories that are reminiscent of those used in codes of alphabet writing. But perhaps because of my field of interest (speech science), it has been persistently clear to me, as a student and researcher, that there are hardly any links between instrumental observations of speech and formal analyses of transcripts. This discrepancy was the source of a career-long interrogation on how it was that empirical observations did not serve to correct assumptions shared by analysts and investigators of spoken-language processing. In the meantime, I became acquainted with a body of historical essays, including a publication by Linell (1982/2005) entitled The Written Language Bias in Linguistics. These works documented how spoken language came to be studied using text and essentially demonstrated that the formal analysis of language, as currently taught in universities, is conceptually based on orthographic code. The essays exposed an important bias with broad implications, although the implications were not spelled out except by reference to the sociology of literacy and language theory. There was a need, as I saw, for a work that documented the course of the writing bias, the arguments used to claim the existence of orthographic-like units and categories in the brain, and the xv
- 22. consequences for experimental research. More pressingly, there was a need to address the bias by detailing how speech can be studied, not through the prism of one’s writing system, but through instrumental techniques that could identify motor-sensory elements and structures of speech processing. However, I fully recognized the risks of such an endeavor. I realized that exposing a bias across sectors of language science ran the risk of appearing confrontational on all fronts and that readers might not be aware of the accumulating evidence of a basic problem. To avoid such judgments, the monograph had to discuss experimental findings. It had to be made clear that, throughout the history of language science, investigators have explicitly acknowledged a basic discrepancy between theories erected on orthographic concepts and observed processes of spoken language. But it was also a concern that a book that documents a bias in research would constitute a wholly negative enterprise – unless it proposed a way of addressing the problem. The monograph had to show that there is a coherent set of findings that supports an approach to the study of spoken language that does not entail notions of letter- and word-like units as in text. The discussion of this evidence, however, presented yet another risk. In this case, there was the chance that some vital findings could be missed, or else that the evidence would refer to domains of research unfamiliar to some readers. On this problem, the challenge was to remain on topic while assuming a knowledge base by which readers could judge competing proposals. As a compromise, I provide, especially in the latter parts of the book that refer to neuroscience, multiple references to recent surveys and critical reviews from which readers can cull background informa- tion. In short, I fully recognized the risks of submitting a work dealing with the writing bias in language science. However, far outweighing these consider- ations was the prospect of a science that seeks to understand the biological underpinnings of spoken language based on culture-specific constructs of writing. In other words, weighing in the balance was the prospect of pursuing studies of speech processes in a way that made the scientific status of the field appear questionable at its base. The present work aims to address the writing bias through an approach that rests on observable structures of speech. This offers a view of how research may move forward in elaborating biologically plausible accounts of spoken language where speech observations are commensurate with neural processes. The evidence that is marshaled in support of the approach draws mostly from published work, though some pivotal findings are the product of my collabor- ation with colleagues whom I wish to thank. In particular, I am indebted to Boutheina Jemel (Université de Montréal), who designed the image of the book cover, and Annie C. Gilbert (McGill University). Both have had a major influence on my view of the role of neural oscillations in speech processing. Both have convinced me of the value of small laboratories where xvi Preface
- 23. experimentalists can transgress the boundaries of academic disciplines and share expertise. I also gratefully acknowledge the support of members of our team, especially Julien Plante-Hébert and Antonin Rossier-Bisaillon without whom there would have been no time to write. The essential parts of this monograph were developed in answering an invitation from Philippe Martin (Université Paris-Diderot) to deliver a series of conferences in his department, and I am truly grateful for his ongoing encouragement and discussions on prosodic structure. The format of the subject matter that follows benefited from the commentaries of students who attended my courses at the Université de Montréal. Hopefully, the monograph can serve to foster critical thinking in future students and researchers. I also thank Douglas Rideout for revising the text under the pressure of impending deadlines. Finally, in the context where there is considerable controversy in language theory, I wish to express my gratitude to my editor Helen Barton and Cambridge University Press for their open-mindedness and support. xvii Preface
- 25. Introducing a Fundamental Problem of Language Science It is quite difficult to conceptualize how spoken language functions without at once referring to writing. Most readers of the present text began to represent speech with alphabetic signs at about four or five years of age. It is therefore understandable that people with years of training in alphabet writing would view speech as containing combinations of letter-size elements, words, phrases, and sentences, like the units on this page. But it should be recognized that these intuitions are not universal. People who learn writing systems such as Japanese kana or Chinese hànzi characters, for instance, conceptualize units and combinations quite differently, and do not represent letter units, or words as groupings of letters separated by spaces (e.g., Hoosain, 1992; Lin, Anderson, Ku et al., 2011; Packard, 2000). Nevertheless, specialists in various disciplines have come to use alphabetic signs, along with other orthographic units and categories, not only in analyzing different languages, but also in clinical tests and research on the neural under- pinnings of speech processing. For an outsider, this might seem odd. Clearly, not everyone knows alphabet writing. How then could culture-specific con- cepts of letters and words, or categories like consonant, vowel, noun, verb (etc.), serve to analyze different spoken languages, let alone neurobiological processes common to all speakers? Indeed, such applications of orthographic concepts have not gone unchallenged. As documented in this monograph, there is a history of academic work in which authors repeatedly criticize the centrism of analyses that use orthographic units and categories, and many question the face validity of language theories based on such analyses. In considering these theories, one needs to weigh the fact that there is scant evidence – from instrumental observations of speech to brain imaging – validating the view that utterances are processed as sentences containing hierarchical sets of phrases, words, and letter-like units, as conceptualized in alphabet writing and conventional language analysis. Researchers who continue to seek these hierarchical sets in speech or brain responses also face a basic problem in that there are no working definitions of what constitutes a “word,” “phrase,” or “sentence,” except by reference to such marks as spaces in text (Dixon & Aikhenvald, 2002; Haspelmath, 2011). As for the belief that speakers create 1
- 26. sentences using rules that serve to combine given grammatical classes of words, there are conflicting views that are outlined in the present work. For instance, neuroscientists have submitted evidence invalidating the idea that the brain processes words in terms of orthographic-like categories such as noun and verb, which undermines decades of formal syntactic theory (see, e.g., Vigliocco, Vinson, Druks et al., 2011). Such results and the failure to ground writing-induced concepts of language analysis in speech bear disquieting implications for the field of study. Yet, despite a substantial body of critical commentary, few works present findings that address the problem of the “writing bias” in language science. For students in psychology, communication disorders, and language science, introductory texts offer little forewarning that core concepts of language study do not link to observable physical or physiological aspects of spoken language. In fact, one pains to find a textbook that mentions the problem. Some works allude to it, though almost as an aside. For example, in the final pages of his Introduction to Neuropsychology of Spoken Language and Its Disorders, Ingram (2007) offers the following terse critique of the history of processing models that assume “interim” units of representation such as letter-like phon- emes and words: Early models of sentence processing (or production) tended to simply borrow the units of interim representation from linguistic theories (competence models). Subsequently, psycholinguists sought evidence from performance constraints for the psychological reality of these units. However, evidence at the neural processing level for an interim level of linguistic representation is scarce at best. (p. 377) This may have prompted some readers to wonder why such criticism of decades of research appears in the latter pages of a text and not as an introduc- tory warning. Other specialists are just as critical of the prospects that language study might one day link to neuroscience. For example, Poeppel and Embick (2005): In principle, the combined study of language and the brain could have effects in several directions. (1) One possibility is that the study of the brain will reveal aspects of the structure of linguistic knowledge. (2) The other possibility is that language can be used to investigate the nature of computation in the brain. In either case, there is a tacit background assumption: namely that the combined investigation promises to generate progress in one of these two domains. Given the actual current state of research, these two positions – rarely questioned or, for that matter, identified in studies of language and the brain – lack any obvious justification when examined carefully. (p. 1) According to these authors and others (Embick & Poeppel, 2015; Grimaldi, 2012, 2017; Poeppel, 2012), the basic problem is one of ontological incom- mensurability, or the fact that “the fundamental elements of linguistic theory 2 Introducing a Fundamental Problem
- 27. cannot be reduced or matched up with the fundamental biological units identi- fied by neuroscience” (Grimaldi, 2012, p. 3). Actually, this incommensurability extends beyond neuroscience and includes various domains of instrumental observation. Speech scientists have long recognized there are no divisible units in signals that match letter-like phonemes in language analysis (Liberman, Cooper, Shankweiler et al., 1967). In a recent meta-analysis of neurophysio- logical models of speech processing, Skipper, Devlin, and Lametti (2017) concluded that “after decades of research, we still do not know how we perceive speech sounds even though this behavior is fundamental to our ability to use language” (p. 78). However, one notes that the preceding formulation of the issue of onto- logical incommensurability does not call into question the validity of con- ventional language analysis, which is not unusual. Poeppel and Embick (2005) assert that, “If asked what to study to learn about the nature of language, surely one would not send a student to study neuroscience; rather, one might recommend a course in phonetics or phonology or morphology or syntax or semantics or psycholinguistics” (p. 2). The presumption that lin- guistic methods serve to analyze the “nature” of language exemplifies a misapprehension of the roots of the ontological problem in language study. It fails to recognize that the issue of incommensurability can arise precisely because units and categories of language analysis do not reflect natural elements, but cultural constructs that draw from a writing tradition. At the risk of stating the obvious, students who, for instance, learn to analyze the constituents of “sentences” by examining distributions of symbols in tran- scripts, or by performing substitutions and commutations of letters, words, parts of words, phrases (etc.), while using such notions as consonant, vowel, verb, preposition, auxiliary (etc.), are not working with natural units and categories. They are principally working with orthographic concepts of Latin grammar and overlooking entirely the signals and physiological processes involved in vocal communication. There is, in this sense, a misapprehension in claiming that current methods of language analysis serve to understand the nature of spoken language. As documented in this book, such claims overlook a body of work criticizing the orthographic bias in linguistic methods. In introductory works, though, readers are only marginally informed of the problems that arise when culture-specific concepts of writing are used in analyzing spoken language. As a further example, in a popular compendium of brain-imaging research on language, Kemmerer (2015b, pp. 273–274) cautions readers against culture-centric assumptions, noting that there are substantial differ- ences in the way languages “carve up” meaning with words. Then, a vast body of neuroimaging research on the processing of lexical items is dis- cussed, much of which refers to test protocols involving presentations of 3 Introducing a Fundamental Problem
- 28. isolated words. Regarding such methods, there are no cautionary remarks that language groups also carve up verbal expressions differently and that meaningful forms may not reflect units like words in European-style writing and dictionaries. This is not simply a technical matter. It is a decisive conceptual issue, one which also carries ethical implications. Although not widely publicized, the cultural specificity of the word concept has led to debates amongst language pathologists confronted by the problem of how to diagnose “word-finding” deficits for speakers of so-called wordless or polysynthetic languages (discussed later on in this monograph). In these cases, linguistic analysis does not serve to distinguish words from phrases or sentences, and the problem is not limited to little known languages like Inuktitut, Mohawk, Cayuga (and others). It can arguably extend to “isolating” languages representing some of the largest communities of speakers, such as Chinese, which does not conceptualize words as in alphabet writing (cf. e.g., Hockett, 1944, p. 255: “there are no words in Chinese”; Packard, 2000, pp. 16 and infra). Certainly, neuroimaging studies that use visual presentations of words like apple, dog, cup (etc.) provide valuable and even essential information on semantic processes and repre- sentations. And clearly isolated units corresponding to space-divided words in writing are used to name objects or actions (such as proper names and imperatives). But people do not speak to each other using isolated forms like apple, dog, cup, and it is inherently difficult to determine how many “words” there are in basic utterances. For instance, in I’m done, Tom’s gone. Take’m, You’re right (etc.), do ‘m, ‘s, ‘re, constitute words? Are the syllables that contain these units subject-verb phrases (and so forth)? Faced with such definitional issues, some analysts contend that, for the most part, speech may not involve words, or combinations of words as in a text, but formulas which are processed as such (Ambridge & Lieven, 2011; Beckner, Ellis, Blythe et al., 2009; Bybee & Beckner, 2009). Thus, beyond caution- ary remarks on language relativity, there is the fundamental issue of whether concepts linked to writing can serve to guide research on the structures and processes by which spoken language conveys meaning. For scientists from other disciplines, the latter concern may seem so basic as to undermine language study as a science. After all, why has the failure to observe notional orthographic units and categories in sensory signals not led to a reevaluation of conventions of language analysis? Some do not see the problem as relating to a writing bias, but instead suggest the need to refine linguistic concepts using “features” and “morphemes” rather than letter-like phonemes and words (Embick & Poeppel, 2015; Poeppel, 2012). Many also point to a general problem of methodology: language specialists need to develop theories that take into account instrumental observations so as to orient what appears to be a stockpiling of eclectic data. For instance, Grimaldi (2012) 4 Introducing a Fundamental Problem
- 29. notes that there is a basic ontological problem in linking theories of language to observations of neural processes: Despite the impressive amount of neural evidence accumulated until now, the field of research results is fragmented and it is quite difficult to reach a unit of analysis and consensus on the object of study. This frustrating state of the art results in a detrimental reductionism consisting in the practice of associating linguistic computation hypothe- sized at a theoretical level with neurobiological computation. However, these two entities are at the moment ontologically incommensurable. The problem lies in the fact that a theory of language consistent with a range of neurophysiological and neuroimaging techniques of investigation and verifiable through neural data is still lacking. (p. 304) Grimaldi mentions one exception. In his view, the language theorist Jackendoff has developed a formal model of “sentence” processing that attempts to connect linguistic analyses to neurophysiological observations. The proposal, the Parallel Architecture Model of Sentence Processing (Jackendoff, 2007b, 2009, 2017), was principally intended as a response to criticisms that formal theories of sentence generation present static arborescent structures that do not take into account that speech unfolds over time (see Ferreira, 2005). In other words, the model attempted to reconcile an analysis of sentences using static signs on a page with temporal aspects of speech. This is a rather obvious discrepancy between theory and observation, which might lead one to ask why language theories do not generally take into account the time dimension of speech and how this shapes the processing of language. On such issues, Jackendoff’s justification for the model is instructive in defining the ontological incommensurability problem that extends well beyond the issue of static formalisms. In particular, in proposing his model, Jackendoff admitted that he was breaking with a long-standing “mentalist” tradition in language study. By this, he meant that, following Saussure’s (1916/1966) division of langue and parole, language analysts have generally adopted a distinction “between the study of language competence – a speaker’s f-knowledge (‘functional know- ledge’) of language – and performance, the actual processes (viewed computa- tionally or neurally) taking place in the mind/brain that put this f-knowledge to use in speaking and understanding sentences” (2009, p. 27). The mentalist perspective adheres to the Saussurean premise that language is a product of the mind that can be studied separately from the speech medium by analyzing transcripts. Although Jackendoff (2009) suggested that the competence– performance division may have originally been intended as a methodological principle (p. 27), this is certainly not the case historically (as clarified in the present monograph). The mentalist tradition, as Jackendoff notes, has been highly influential and was a principal vector of cognitive psychology and cognitive neuroscience. Indeed, G. Miller (2003) recounts that in the 50s, 5 Introducing a Fundamental Problem
- 30. when behaviorist theories were being overturned by a group of authors for which he was a spokesman, he hesitated to use the term mentalist to describe the views of the group and instead referred to a “cognitivist” approach (p. 142). For Jackendoff, the competence–performance division in the mentalist trad- ition had some unfortunate consequences in that some theorists “tended to harden the distinction into a firewall: competence theories came to be con- sidered immune to evidence from performance” (2009, p. 28). But this is not the only break with the mentalist tradition that Jackendoff requests for his proposal. To respond to the lack of agreement between formal language theor- ies and neuroscience, he suggests a reassessment of a modular “syntactico- centric” view which has dominated language theory for over half a century and which, in his opinion, was “a scientific mistake” (p. 35). In the end, the Parallel Architecture Model attempts to answer the problem that speech unfolds over time. Even so, the model operates on conventional units and categories where sentences are seen as hierarchical assemblies of letter symbols, words, phrases (etc.), all of which have no general physical or physiological attributes in speech. Again, this incommensurability is not seen as particularly troublesome for the theorist who accepts that units, such as words, “as is well known, are not present in any direct way in the signal” (Jackendoff, 2007a, p. 378), or that letter-like phonemes are not in the “physical world” (Jackendoff, 2017, p. 186). But if such units are not in signals, in physical manifestations of spoken language, then how does neurophysiology operate to extract and process phonemes and word combinations? How do scientists test putative grammars, or how would a child acquire a combinatorial scheme if units are not present in sensory signals? There is in such views a “firewall” of sorts preventing any invalidation of the writing-induced con- cepts of language analysis and theory. Thus, while some authors acknowledge the historical failure to connect the study of language competence to perform- ance (which includes physiology and signals), the presumption that spoken language can be studied using transcripts and orthographic concepts remains. This has led to enduring pseudo-puzzles that extend to debates on the origins of spoken language. For instance, how would formal grammars based on linguistic analyses emerge in human biology if the units on which the gram- mars function were not in the physical medium of communication? On such pseudo-puzzles, the failure to ground conventional units of language descrip- tion in sensory signals has contributed to speculations of their innateness and saltations in the evolution of the mind/brain that present logical problems (as outlined by Christiansen & Chater, 2008). On the other hand, rather than addressing the problem of using writing concepts in the study of spoken language, some ask if the concepts are not “unavoidable descriptive conveniences” (Bybee & McClelland, 2005; Jackendoff, 2007a, p. 352). Certainly, examining transcripts and orthographic 6 Introducing a Fundamental Problem
- 31. units can provide useful information in several areas of inquiry, especially in sectors relating to reading. Nonetheless, static writing signs do not represent speech acts and signals. No matter how fine the transcripts are, they do not capture such things as muscle contractions, sound properties, breath flow, or any other physical or physiological aspect of verbal communication. Thus, writing signs offer no means by which to explain the structure of spoken language. As a consequence, researchers who refer to linguistic analyses and theory will not find a definition of the nature of units such as letter-like phonemes, words, phrases, or sentences. These forms used to describe spoken language are generally taken as given in sectors of language study, but never explained. More importantly, viewing spoken language through writing over- looks the multisensory context of speech that essentially defines the function and meaning of utterances. This latter problem is commonly acknowledged: one cannot interpret the meaning of utterances, let alone understand how they convey meaning out of context. Yet current models of language are based on the analysis of script that completely removes the object of study from the motor-sensory medium and communicative environment. The following monograph presents evidence, some of which has been available for some time, showing that one does not need to refer to writing concepts to study the processes of spoken language. In discussing this evi- dence, it is a contention that postulates of interim units and categories of language analysis do not exist beyond “descriptive conveniences” for obser- vers who know alphabet writing, and are not “unavoidable” as some have argued. There are observable structures of speech that are universally present across languages. All spoken languages present syllable-like cycles, prosodic groupings, and units of speech breathing. It is also the aim of this book to show that findings linking these structures to neural processes present a major shift in the way semantic representations are conceptualized. Many readers who have formal training in language study are likely to object to such a viewpoint based on the notion that language is separate from speech. In the context where this belief is pervasive, any attempt to address the problem of the incommen- surability between conventions of language analysis and observations of speech processes requires a critical look at the historical arguments that have served to maintain a speech–language division. This also extends to a division in methods of inquiry where linguistic analysis is often seen to have theoretical precedence over instrumental observations. These issues have essentially guided the organization of the subject matter of the present monograph into four parts. Part I, entitled “Questions of Ontology: Writing and the Speech–Language Divide,” documents the source of the belief that language is separate from speech. For many historians, this belief originates in the practice, instituted by philologists in the nineteenth century, of viewing spoken language via script. 7 Introducing a Fundamental Problem
- 32. Influential authors such as Saussure (1857–1913) formulated pivotal argu- ments for separating langue and parole that are still echoed in textbooks. In reviewing these arguments, an essential point of Part I is that early language theorists had no instruments by which to record and visualize speech. When instrumental methods became available, the notion that one could work out feature systems by examining letter signs on a page was already accepted in schools of phonology. But a turning point occurred when early instrumentalists reported that writing-induced concepts of linguistic methods did not reflect in speech. At that point, the argument was made that phonological criteria were essential in orienting observations. Many interpreted this to mean that empir- ical research had to be hypothesis-driven in terms of the assumed units and categories of linguistic theory. Part II, “Questions of Epistemology: The Role of Instrumental Observations,” examines the consequences of the idea of the primacy of linguistic-type analyses in guiding research. It is the case that, in using instru- mental techniques, investigators need to have some idea of what exactly in speech serves a communicative function. However, such considerations offer no justification for the belief that research needs to be driven by orthographic concepts of linguistic descriptions. Yet this presumption prevails in sectors of language study. One key example discussed in Part II is the perennial debate on the existence of letter-like phonemes. Despite the acknowledged absence of these units in speech, it has been claimed that types of data, such as spoonerisms transcribed with letters, confirm the existence of phonemes. The assumed primacy of these types of indirect observations over instrumental evidence reflects an epistemological problem and a writing bias that has broad consequences. For instance, some investigators critically refer to studies of spoonerisms in arguing that phonemes are part of an innate competence underlying alphabet writing, a view that has been severely criticized. Part II reviews a body of evidence confronting the belief that language is separate from speech and can be studied using concepts of writing. Of course, not all authors share this viewpoint. Some, in fact, explicitly reject the speech– language division. This entails a major epistemological shift in that abandoning the division implies that structures that are readily identified in motor and sensory aspects of speech are the actual structures of spoken language. Part III, “The Structure of Speech Acts,” and Part IV, “The Processing of Speech Meaning,” address the problem of the ontological incommensurability between conventional linguistic analysis and observations of spoken language. Part III illustrates how one can study spoken language in context where, instead of focusing on script, one refers to the structural attributes of utterances. Viewed as speech acts, utterances are much like other bodily actions that are performed in a changing environment except that the actions modulate air pressure for communicative purposes. As already mentioned, the structure of 8 Introducing a Fundamental Problem
- 33. these actions is reflected in syllable-like cycles, temporal chunks, and breath units, all of which emerge from circumscribable processes of motor speech. A central thesis of the present book is that multisensory context information binds to utterance structure, more specifically, to chunks of articulated sounds via mechanisms of motor-sensory coupling and neural entrainment. On this basis, episodic and semantic memory of sensory experiences that accompany utterances can link to action chunks that some authors view as verbal formulas and semantic schemas. The relevance of this approach is discussed in the context of ongoing problems that arise in models of speech and semantic theory that attempt to provide an interface between neural processes and writing- induced concepts of language description. In sum, Parts I and II offer a review of historical developments underlying a writing bias that has created fundamental problems in research on spoken language, whereas Parts III and IV present an approach that serves to address these problems. However, the background information provided in the first parts should not detract from the main subject of the present monograph. This is not a book about writing, although it refers to critical works on the influence of what some historians call scriptism in language study. It is a book about language science and how one can view processes of spoken language without reference to culture-specific concepts of writing. The latter parts of the mono- graph develop this view in terms of a body of research findings. This research sheds light on how memory of multisensory experiences binds to structures of motor speech and how activations of semantic or episodic memory in relation to these structures underlies a context-based interpretation of utterances. However, given that many may believe that speech is functionally separate from language and semantics, it is useful to begin by examining how this belief, and conventional concepts of language analysis, arose from a tradition of viewing spoken language through script. 9 Introducing a Fundamental Problem
- 35. Part I Questions of Ontology: Writing and the Speech–Language Divide
- 37. 1 How We Are Introduced to the Study of Spoken Language For many readers of the present work, the notion that language is separate from speech (or that language competence is separate from performance) will likely have been acquired in an introductory course to language study. This notion can have a major influence on how language is conceptualized, and presents a central tenet in the field. In examining introductory texts used by generations of students (e.g., the multiple editions of Akmajian, Demers, Farmer et al., 2010; Fromkin, Rodman, & Hyams, 2013, and others), one finds typical arguments for distinguishing speech from language. These arguments serve to specify not only the object of study, but also how to study it. However, all these arguments, it should be pointed out once more, involve written material. For instance, in sections pertaining to the “structure of language,” examples of sentences or words are usually presented in regular orthographic form or tran- scribed using letter symbols of the IPA (International Phonetic Alphabet). These can be used to illustrate that sound–meaning associations are arbitrary (sym- bolic), and that the way sounds are combined to form words, or the way words are assembled to form phrases and sentences, reflects knowledge of combinator- ial rules. The implication of such examples is that this has nothing to do with the vocal processes of speech (or any modality of expression for that matter). The illustrations can captivate an audience, partly because they cater to received notions and intuitions. Of course, after decades of training in alphabet writing, the audience will “know” units like words and sentences, orthographic rules of letter and word combinations, as well as the categories on which the rules apply. On the other hand, one is likely to find some ambiguity on how linguistic descriptions relate to physical and motor-sensory aspects of speech communica- tion. Although textbooks generally contain a section on the “sounds of lan- guage,” many do not contain a single illustration of acoustic waveforms and simply refer to IPA signs. In those textbooks that do contain oscillograms or spectrograms (e.g., Akmajian et al., 2010), there is no indication that it is impossible to divide signals in letter-like units without creating noise, or that “distinctive features” of sounds are not bundled together as IPA letters imply. The misrepresentation in this case is that, in describing spoken language with writing symbols, it is held that one is representing the sounds of a language when, in fact, 13
- 38. the symbols do not reflect signals, suggesting instead the existence of abstract elements. Invariably, the presentations lead to a view of spoken language as reflecting a mental ability that organizes symbolic elements of various orders (at the phonological, morpho-syntactic, and semantic levels of description). Such an ability or “competence,” as the arguments run, is not a matter of the speech apparatus (usually presented by static sagittal representations of the oral and nasal cavities), but constitutes a separate capacity in the brain. Given that this capacity is seen to operate on strings of symbolic units, then the tacit assumption is that symbols of alphabet script can be used to study it. Presenting the object of study this way and focusing on text material can foster the conviction that what defines spoken language is its combinatorial or syntactic character. Thus, to borrow a term from Jackendoff (2009), language study appears, in its introductory ontology and epistemology, to be “syntactico- centric.” The examples that are provided can also create the belief that processes governing communication can be studied using familiar writing codes, without any immediate need to refer to principles of physics and physiology. Yet intro- ductions such as these, which centrally involve examples based on transcripts, distort the role of spoken-language communication in rather obvious ways. Compared to the presentation of written material, one can consider how the use of audio or audiovisual examples of actual speaker–hearer interaction would lead to a very different concept of spoken language and how to study it. This would be especially clear in a case where one is confronted with interactions involving unfamiliar languages. 1.1 Language as an “Autonomous” System, or the Effects of Scriptism Anyone attempting to analyze heard utterances in an unfamiliar language will quickly realize that this in not at all like analyzing one’s own language. Upon hearing unfamiliar speech, a listener can perceive structured sounds along with situational- and speaker-related information. But one would not “know” where to divide words, phrases, or sentences, nor be able to comprehend the utterance. Indeed, intuitions of orthographic units and categories can inherently bias first- hand observations. One revealing example is discussed by Gil (2002) who, in his fieldwork, was confronted with unfamiliar South Asian languages. Consider a situation where an analyst, who does not speak Riau Indonesian, hears an utterance in that language and then proceeds, as Gil does, to transcribe the sounds as “ayam makan.” If one reduces what is heard to this script, then the actual patterns of sounds including syllable-like cycles, groupings, and tonal contours need to be inferred by a reader of an alphabet. Thus, depending on the experience of the reader, it may not be clear what the space and letter sequences actually represent in terms of sound patterns, or if the original patterns heard 14 Questions of Ontology
- 39. by the analyst would be better represented as “ay amma kan,” “ayam mak an,” “a yam makan,” (etc.). Of course, in recording the speech event with alphabet symbols, most of the structural and all of the situational attributes of sounds relating to the speaker’s identity, affect, state (etc.) and the general utterance context are omitted. The idea that an observer can skip over these structural and situational details and analyze content in terms of syntactic-semantic units led Gil to the following graphic analysis of constituents, which, as such, offers little if any information as to how the utterance conveys meaning. EvenifoneweretoconsultadictionaryofIndonesiantotranslatetheconstituents into English, one could end up with nonsensical versions. (In this case, Gil offers several possible interpretations having to do with chickens and eating or being eaten.) In reality, it is only by considering what is occurring in the context of speech that an observer can interpret the signals, their structure, and their expressive character, so as to understand the meaning of the sounds. Hearing other situated utterances would ultimately serve to discern groupings and infer their function – inasmuch as the meaning of utterances can be ascertained from a communicative setting,atleastininitialanalyses.Buttheideathatonecanunderstandhowaspoken language works by analyzing the syntax of alphabet symbols on a page is an illusion. Such impressions arise when interpreting transcripts that represent a familiar language, where the function, structure, and meaning of sounds have been acquired through extensive experience in speaking the language. Although this may seem obvious, it is a common oversight in textbooks and is central to understanding the persistence of the problem of the ontological incommensurability between language theory and instrumental observations, as described in the Introduction. The tradition of analyzing utterances by way of transcripts and orthographic concepts fosters a view of spoken language as an autonomous combinatory system that operates on the very concepts that are used in the analysis. In this looking-glass approach, spoken language appears as an inherently mental function that assembles given syntactic-semantic units, all of which seem to have little to do with the motor and sensory processes of communication. This idea of “language autonomy,” which tends to overlook the biasing effects of orthographic constructs, has led to perennial debates on the grounding of language theory (for polemics on the autonomy principle, see Derwing, 1980; Lyons, 1991; Newmeyer, 1986a, 1986b, 2010; Taylor, 2007, among others). To characterize the problem briefly, an autonomy principle holds 15 How We Are Introduced to the Study of Spoken Language
- 40. that a theory of language can be elaborated using linguistic-type analyses, without reference to other types of data. Newmeyer (2010) illustrates this principle using a chess-game analogy that, interestingly, is the same analogy used over a century ago by Saussure (1916/1966) to illustrate his assumption of a division between language and speech. As Newmeyer sees it: Chess is an autonomous system: There exists a finite number of discrete statements and rules. Given the layout of the board, the pieces and the moves, one can “generate” all of the possible games of chess. But functional considerations went into the design of the system, namely to make it a satisfying pastime. And external factors can change the system. Furthermore, in any game of chess, the moves are subject to the conscious will of the players, just as any act of speaking is subject to the conscious decision of the speaker. So chess is both autonomous and explained functionally. (p. 4) That is to say, knowledge of the combinatorial rules of language is the central object, not the instances where a speaker forms particular utterances during functional acts of communication, or changes in the language code. The key term in this analogy is “Given ... .” In analyzing a familiar language using transcripts where units and categories of an orthographic system are taken as given, or assumed to be known by all, one could formalize a set of rules that generate combinations. This has been the program of some linguists who have elaborated formal models of “sentence” generation said to constitute a “universal grammar.” But then influential models have often been based on analyses of English sentences, usually represented in regular orthographic script (an issue further discussed in Chapter 3). This brings forth obvious questions. In particular, it leads one to ask whether combinatorial models derived from these analyses of sentences can formalize actual operations on elements in the brain or if they reflect artifacts of orthographic conventions. As mentioned in the Introduction, there is at this time no physiological evidence supporting the type of hierarchical combinations of letter-like phonemes, words, phrases, and sentences, which are the hallmark of formal language analysis and models. Although the absence of evidence does not constitute invalidation, view- ing studies such as Gil’s (2002) brings to light the difficulties of an approach where it is assumed that units and categories of Latin grammar are given and known by all. Gil could not presume such elements for Riau Indonesian. He opted for an ad hoc category “S” in the previously cited example because, contrary to familiar European languages, syntactic elements in Riau can be distributed just about anywhere in an utterance and have no case markings. In considering such instances, one can wonder what formal theories would look like if, instead of constructing a grammar based on analyses of English sentences, one took Indonesian as a familiar starting point. Gil’s study, entitled Escaping Eurocentrism: Fieldwork as a Process of Unlearning, suggested the need to question received notions in analyzing languages. In fact, numerous reports 16 Questions of Ontology
- 41. attesting to a diversity of categories and units tend to undermine the claimed universality of some formal theories (see, e.g., Evans & Levinson, 2009). This does not mean that there are no universal processes, nor is the problem a matter of “language relativity.” There are, as discussed in the present work, elements and structures that arise in all spoken languages as a result of constraints on motor and sensory processes of vocal communication. However, it has to be asked if analyses of transcripts are at all suited to the study of these processes, or in guiding research on how spoken language conveys meaning. To illustrate this last point, consider a simple response like “hello” heard in the context of a phone call. Linguistic analysis hardly serves to understand how the sounds can be variably interpreted. Yet upon hearing the sounds, listeners can readily identify a familiar voice, a speaker’s dialect, affective state (etc.), along with coded indices, subtle features that are part of “knowing the language” and that indicate whether the speaker is receptive to the call, is happy to hear from an old friend, and so forth. There is no formal repertoire, no “grammar” or “dictionary” of such speech attributes. But the fact that they are used to interpret utterances in various ways implies a coding of these features in memory. The variety of sound features and associations that can be formed vastly outnumber IPA signs. Yet these attributes are not considered in linguistic analyses, principally because the ana- lyses are limited to what can be represented in writing. Recognizing this aspect of linguistic methods is central. It helps to understand how the tradition of using transcripts has led to a division of the object of study where “language” is abstracted away from the sensory and motor processes of oral communication. Viewing utterances through a known writing system, the object appears to function in terms of conceptual units and rules that are difficult to dissociate from the analyst’s knowledge of a writing code. Several authors of historical essays have designated this effect, and the tendency to conceptualize spoken language through writing, “scriptism” (Harris, 1980, and see Coulmas, 1989, 1996, 2003; Harris, 1990, 2000; Linell, 2005; Love, 2014; Olson, 1993, 2017). As documented in these studies, scriptism not only refers to the particular biasing effect of alphabetic concepts, but refers more generally to the fallacy of claims that analyses of transcripts serve to understand spoken language. To the contrary, the aforementioned essays illustrate that, across cultures, “writing systems provide the concepts and categories for thinking about the structure of spoken language rather than the reverse” (Olson, 1993, p. 2). The problem is that structural attributes of utterances are not determined by orthographic representations of sentences. They are determined rather by motor-sensory processes of oral communication that, obviously, cannot be studied via script. As will be noted in Part II of this book, the response from some authors to the preceding criticism has been to argue that the concepts of alphabet writing are special in that they can capture the true units of a biological competence for language (see also Wolf & Love, 1997). In the end, the arguments imply a view 17 How We Are Introduced to the Study of Spoken Language
- 42. on the superiority of European systems, despite the avowed absence of evi- dence for alphabet-style units in signals and neural responses. However, these arguments largely deal with a side issue to scriptism. They do not address the basic, ontological difference between sentences on a page and utterances. Using any writing system to record spoken language as such disembodies and decontextualizes the object of study. It creates an object that is ontologic- ally separate from observable aspects of spoken-language communication. In summary, the textbook assumption that linguistics examines the sounds of spoken language cloaks the fact that language analysis, as currently taught and practiced, largely focuses on script. This tradition has led to a conceptualization of spoken language as an autonomous system operating on letters, words, and sentences separate from observable structures of utterances. As is known, the autonomy concept was formally introduced in language theory in the nineteenth century. Theorists like Saussure formulated influential arguments in claiming that language is separate from speech and reflects a faculty in cortical areas of the brain. Similar arguments, including Saussure’s chess-game analogy, can be found in textbooks where the idea of a separate language faculty has been reworded as a competence–performance division. Historically, this division was central to the rise of linguistics as a discipline. But it is also the locus of a fundamental problem facing research. There is, as many recognize, a basic incompatibility between theories erected on linguistic analyses and observations of speech. That this problem persists speaks to the influence of foundational works. In reviewing the arguments used by Saussure, the following sections draw attention to the context in which they were formulated. Specifically, it should be acknowledged that language theorists in the nineteenth century could not record, let alone visualize, sound patterns of utterances, and the idea of cortical faculties was dominant at the time. However, the concept of localized faculties has long since been abandoned. In light of available knowledge about the brain and current recording techniques, language study in its focus on script and reference to separate faculties seems oddly out of step. Perhaps the clearest example of this nineteenth-century viewpoint is the way in which “speech” is separated from language in introductory texts, as illustrated below. 1.2 Defining “Speech” In fields of study bearing on spoken language, including several applied sectors, speech is typically presented in terms of a vocal production of sounds and is traditionally modeled as a source-filter function (e.g., Fant, 1960). It needs to be recognized, however, that speech cannot be defined this way precisely because one can create countless vocal resonances, such as in a source-filter system, without producing speech. In other words, it is inherently difficult to separate speech from language. Thus, “speaking to someone” is commonly understood as 18 Questions of Ontology
- 43. a communicative act involving a language code (although not all information in speech is coded). Some cultures have terms to designate both language and speech without implying separate processes (e.g., eine Sprache in German, or un parler in French). But this is not simply a question of terminology. In areas of language study, mechanisms of oral expression are traditionally presented as if they could be physiologically isolated from language functions, usually localized in cortical areas. Examples of this can be found in practically all textbooks where “speech” is illustrated using schematic midsagittal sections, such as in Figure 1.1(a). Such representations have been used by analysts and teachers since the nineteenth century, and can even be found in works of neuroanatomy (e.g., Scott, Wylezinska, Birch et al., 2014). The reality is that illustrations of anatomical structures like (a), which are meant to refer to articulatory and phonatory systems, do not as such produce speech. Clearly, processes of the central nervous system (CNS) are involved, which go beyond a source-filter concept. In this light, some might argue that brainstem structures and other subcortical processes may provide a basis for separating speech functions from language and, in fact, this division has been applied in clinical sectors where “speech and language” disorders have been traditionally viewed as involving subcor- tical and cortical lesions, respectively. Thus, one might use a drawing like Figure 1.1(b) to suggest a speech–language division where some subcortical structures are assigned to speech. But as to whether this anencephalic-like concept might more aptly illustrate a speech–language division, one can refer to actual cases of anencephaly, as in (c). Anencephalic infants do not develop a telencephalon (and many do not develop parts of the diencephalon). Yet some have survived beyond two years of age and thus beyond the stage where normal infants begin to babble (at about 6–8 months; Dickman, Fletke, & Redfern, 2016; Oller, 2000). In rare instances, the infants have a relatively intact brainstem along with neural pathways and structures supporting vision and hearing. In such cases, it has been reported that many behaviors traditionally ascribed to cortical activity are present and can originate in the brainstem (Shewmon, 1988). For instance, these infants can react to noxious food and produce avoidance responses including vocalizations and cries. They show interactions with moving objects involving eye movements accompanied by head turning (indicating a degree of motor-sensory coupling), and also produce facial expressions including smiles (for a recent review of studies of anencephaly, see Radford, Taylor, Hall et al., 2019). But they do not babble, or produce anything resembling speech (i.e., they do not develop eine Sprache, un parler). The point is that, contrary to a Saussurean ontology, speech is not a separable physiological function and cannot be defined as a passive source-filter system. And just as problematic is the idea that language is separate from speech. 19 How We Are Introduced to the Study of Spoken Language
- 44. 1.3 Was the Speech–Language Division Ever Physiologically Grounded? Some may contend, as did Jackendoff (2009, p. 29), that the separation of speech from language or the performance–competence division was simply intended as a methodological or heuristic principle. This is not the case. Early in the history of language study, Saussure formulated a set of tenets for language analysts based on the belief that the speech–language division was physiologically attested and that there was a separate “language faculty” in the brain. This viewpoint has had a lasting influence. As R. Harris (2002) noted: “The basis of linguistic theory has remained in all essentials unchanged since it was first laid down in Saussure’s Figure 1.1 Typical schematic sagittal representation of the “speech” apparatus. (a) Usual representation as can be found in introductory texts to language study; (b) with added brainstem and cerebellum for an anencephalic version of the sagittal schematic. (c) Photo of an anencephalic infant; see discussion in the text (from Bathnagar, 2012, with permission). 20 Questions of Ontology
- 45. Geneva lectures of 1907–11” (p. 21). The speech–language division is a prevailing doctrine, and also reflects nineteenth-century views of a hierarchical brain (Parvizi, 2009). However, the original claim of a physiological division no longer holds, and arguments for the separate study of language have been criticized on fundamental grounds. 1.3.1 Saussure’s Argument of a Separate Language Faculty in Broca’s Area In terms of neurophysiological evidence, at the time of his Cours de linguis- tique générale, Saussure believed that the capacity to create a system of signs was located in Broca’s area (Broca, 1861). In discussing this function, Saussure distinguished language from the motor-control processes of parole, though he did not attempt to localize the latter. In citing Saussure’s Cours, it is worth mentioning that the English translation by W. Baskin alters key terms in that “speech” is often used where Saussure uses langage, so the following excerpt restitutes the term “language” (in brackets) that is in the French version of the Cours. It is perhaps useful to note that compared to English, which has two terms to designate language and speech, French has four: parole, langue, langage, and un parler. In Saussure’s view, language (langage) is the capacity to constitute a system of coded signs that can be a spoken code or langue, a writing code, or other systems of signs including gestural “sign languages.” Saussure’s central claim is that the division between speech and language is not merely methodological, but physiological: We can say that what is natural to mankind is not oral speech but the faculty of constructing a language, i.e. a system of distinct signs corresponding to distinct ideas. [. . .] Broca discovered that the faculty of [language] is localized in the third left frontal convolution; his discovery has been used to substantiate the attribution of a natural quality to [language]. But we know that the same part of the brain is the center of everything that has to do with [language], including writing. The preceding statements, together with observations that have been made in different cases of aphasia resulting from lesion of the centers of localization, seem to indicate: (1) that the various disorders of oral [language] are bound up in a hundred ways with those of written [language]; and (2) that what is lost in all cases of aphasia or agraphia is less the faculty of producing a given sound or writing a given sign than the ability to evoke by means of an instrument, regardless of what it is, the signs of a regular system of [language]. The obvious implication is that beyond the functioning of the various organs there exists a more general faculty which governs signs and which would be the linguistic faculty proper. (Saussure, 1916/1966, pp. 10–11) The idea of a separate language faculty is still present in the literature. Yet setting aside the historical debate on localizing aphasia (see, e.g., Ross, 2010), it has been recognized that the notion of a circumscribable language function in 21 How We Are Introduced to the Study of Spoken Language
- 46. Broca’s area is incompatible with clinical evidence, some of which dates back to the 50s. One early demonstration was provided by Penfield and Roberts (1959). These authors surgically removed large parts of Broca’s and neighboring areas in an individual who, following surgery, showed minor effects on spoken language that dissipated within a few weeks (although, on p. 163 of their report, the authors mention that the individual presented congenital problems that could have caused a “displacement of the function”). With the exception of a report by Marie (1906), these authors were amongst the first to suggest that processes of spoken language involve subcortical structures traditionally associated with “motor-control” pathologies: “Since all these removals of the convolutions that surround the speech areas do not produce aphasia, it seems reasonable to conclude that the functional integration of these areas must depend upon their connection with some common subcortical structure” (Penfield & Roberts, 1959, p. 212). Current understanding and clinical evidence largely support this conclu- sion. As Murdoch (2010a) notes: “Clinical evidence is now available to show that permanent loss of language does not occur without subcortical damage, even when Broca’s and Wernicke’s areas have been destroyed by lesions” (p. 79). Adding to these clinical observations, case reports of children who have undergone the excision of their left or right cortex, but who nonetheless develop spoken language, also undermine the thesis of a localized language faculty in Broca’s area (Bishop, 1983; Curtiss & de Bode, 2003). Such findings are part of a body of work that challenges a cortico-centric conception of spoken language where language is essentially seen to involve cortico-cortical couplings with white-matter connections to subcortical structures subserving motor functions (see the critique of Skipper, 2015; Skipper et al., 2017). This hierarchical view of the brain has changed. There is a now a consensus on the point that motor- related speech dysfunctions associated with structures like the basal ganglia and cerebellum accompany semantic-related deficits, and that aphasia as such can arise from lesions in the thalamus (Crosson & Haaland, 2003; Murdoch, 2001; Murdoch & Whelan, 2009). The findings have major implications with respect to a Cartesian concept of language as separate from motor-sensory functions. In extending this changing concept of spoken language, the present monograph refers to a body of evidence that illustrates how constraints on motor-sensory process of speech can shape structures of spoken-language processing. The point is that, in light of current knowledge, the doctrine of a speech–language distinction is neither physiologically based nor methodologically useful. On the contrary, it has hin- dered any consideration of the inherent links between motor-sensory and semantic functions, as further documented in this work. For language theory and analysis, the previously mentioned set of findings points to fundamental issues that have yet to be fully weighed in sectors of language study. In the foundational principles of his Cours, Saussure assumed 22 Questions of Ontology
- 47. that spoken language involves an autonomous system located in Broca’s area, and saw this system as physiologically separate from speech. Based on these assumptions, he proposed methods by which to analyze the abstract system using transcripts as he saw no need to refer to the structural aspects of motor-sensory processes of vocal expression. To the contrary, research shows that there is no cortically localized language function, and that “language” and motor-sensory processes intertwine. There are several impli- cations for Saussure’s approach to the study of spoken language via tran- scripts, one of which should be clear: if there is no physiological basis for dividing language from speech, then observable structures of utterances are the structures of spoken language. In fact, the natural structures of spoken language were never “given” or represented by culture-specific writing signs and units. In continuing to idealize language via transcripts and concepts relating to letters, words, sentences (etc.), analysts fail to recognize that the principal historical justification for using script in the study of spoken lan- guage was the nineteenth-century view of an autonomous language faculty, which no longer holds. 1.3.2 Arguments of the Arbitrariness of Signs and Abstract Phonology Saussure presented several influential arguments supporting a Cartesian view of language that still prevails. In particular, Saussure taught in his Cours that the capacity to create arbitrary (symbolic) sound–meaning associations and phonological systems demonstrates that a language function is separate from an oral modality of communication and is solely governed by what he called “rational” principles (which can be seen to refer to cognitive functions before G. Miller formally introduced the term). The linguistic sign is arbitrary; language, as defined, would therefore seem to be a system which, because it depends solely on a rational principle, is free and can be organized at will. (p. 78) [. . .] Just what phonational movements could accomplish if language did not exist is not clear; but they do not constitute language, and even after we have explained all the movements of the vocal apparatus necessary for the production of each auditory impres- sion, we have in no way illuminated the problem of language. It is a system based on the mental opposition of auditory impressions, just as a tapestry is a work of art produced by the visual oppositions of threads of different colors; the important thing in analysis is the role of the oppositions, not the process through which the colors were obtained. (p. 33) The quoted assertions are characteristic of the Cours. It is the case that signs in spoken language appear arbitrary and that distinctive features vary across lan- guages. But the conclusion does not follow that signs and feature systems are 23 How We Are Introduced to the Study of Spoken Language
- 48. mental constructions that “depend solely on a rational principle” and are “free and can be organized at will.” Certainly, spoken languages can draw from a variety of sound features that can be produced. And a speaker may variably associate sounds and meaning. Yet neither features nor arbitrary signs arise independently of the possibilities afforded by a motor-sensory medium of expression. In omitting this point, Saussure’s arguments have two major consequences on theories that adopt the doctrine of a speech–language divide. First, viewing language as a mental construction foregoes the need to explain how arbitrary signs and distinctive features arise. These are simply taken as given, or the product of a given faculty. A second, and more important consequence of the doctrine is that it excludes, de facto, any account where constraints on the medium of expression are viewed as factors that can shape processes and signs of spoken language. This is clear in Saussure’s outright denial that mechanisms of production and perception have anything to do with language. Such a theoretical fiat, in effect, denies some rather obvious links between structural attributes of utterances and constraints on the vocal modality – which are readily observed inasmuch as one examines utterances as signals and not as sentences on a page. For instance, no “rational” principle or mental exigency forces the linear arrangement of sounds and verbal elements in utterances. This is not a product of thoughts or sensory experiences. When one sees or imagines someone jumping, one does not first think of a person then an action, or an action separate from the individual performing it, as when saying “Paul jumps.” This linear formatting of sounds and associated meanings is imposed by the medium, by the fact that the sounds in utterances unfold over time. The medium of articulated sounds also imposes structure in the forms that are linearly arranged, as illustrated in the acoustic waves of Figure 1.2. In viewing the patterns in this figure, one might ask what possible mental or “rational” principle requires the cyclical closing and opening motions of the vocal tract to create syllable-like pulses of air, harmonic patterns, alternating long and short cycles marking rhythms, or shifts in fundamental frequency (F0) that accompany breath-divided utterances. These observable structures and units can be explained by reference to mechanisms of the motor-sensory medium (as detailed in Part III of this book), and serve spoken-language communication. But they cannot be understood as products of the mind. As for the arbitrariness of signs and the “symbolic” aspect of spoken language, Saussure did not consider how the uniquely human capacity for symbolic language links to an equally unique ability to produce and control orally segmented vocalizations (a point developed in Chapter 2). This is understandable given the state of research in comparative anatomy at the turn of the twentieth century. But the issue here is how such observations continue to be overlooked under the assumption of a speech–language division. There are historical reasons for the persistence of this viewpoint, which has created 24 Questions of Ontology
- 49. a firewall against invalidating evidence, as Jackendoff (2009) has remarked. One critical point of disconnect between theoretical concepts of language analysis and empirical observations occurred with the rise of phonology. 1.3.3 On the Primacy of Linguistic Criteria: The Historical Disconnect from Instrumental Observations In considering language analysis, one might wonder why readily observed structures, as visualized in Figure 1.2 and which are universal in speech, are not generally viewed as “design features” of spoken language (Hockett, 1960). A principal historical reason for this oversight may be that the structural patterns are not easily transcribed. On this point, it is essential to note that, when methods of linguistic analysis were being devised, the basic instruments used today to record and view speech signals were not widely available. Saussure, for instance, was a philologist who worked on ancient texts and lectured on the reconstruction of proto-Indo-European. At the time, transcripts were the main data if not the only data on which to formulate an understanding of spoken language. Instruments that could visually record speech, such as the kymograph (Vierordt & Ludwig, 1855), were rarely used, and the first spectro- graph only appeared decades later (Steinberg, 1934). But there was also some reticence as to the usefulness of instrumental observations, even amongst classical phoneticians. The reason was that both philologists and phoneticians were applying subjective classifications of speech sounds and many did not see the need for instruments (e.g., in 1913, Jespersen saw that “Les instruments qui permettent de mesurer les différences de position des organes et de représenter par des chiffres la tension des muscles ou l’amplitude des ondes [. . .] sont inutiles”; in Auroux, 2000, p. 511; and see Malmberg, 1972). A similar disin- terest for instrumental methods prevailed amongst early phonologists of the Figure 1.2 Observable acoustic structures of an utterance: Cyclical patterns in energy contours, temporal groupings, tonal contours, and both amplitude and pitch declination across breath units of speech. 25 How We Are Introduced to the Study of Spoken Language
- 50. Prague School in the 1920s, which included such influential authors as Trubetzkoy (1890–1938), Jakobson (1896–1982), and adherents like Martinet (1908–1999), Benveniste (1902–1976), among others (see Fischer-Jørgensen, 1975). How, then, did analysts come to conceptualize systems of sound oppos- ition as put forward by Saussure without sound-recording instruments? A reading of early works offers a view of how the analyses were performed. At the time of Saussure’s Cours, philologists were engaged in comparing languages often using ad hoc alphabet-style symbols. With a view to reconstructing proto- Indo-European, analysts of ancient texts focused on changes in sound distinc- tions, which were described using impressionistic articulatory categories. Terms like “stop,” “fricative,” “front,” “back,” “palatal” (etc.) were broadly used by both philologists and phoneticians like Passy (1859–1940) and Sweet (1845– 1912). But it is worth noting that these categories were subjectively established. Thus, when classical phoneticians categorized sounds according to articulatory place and mode, they did so by reference to what they felt they were doing when producing the sounds. Moreover, in an effort to standardize descriptions, Sweet (1877) suggested that different signs should be used only for “those broader distinctions of sound which actually correspond to distinctions of meaning in a given language” (p. 103), although this was also problematic. To illustrate the way speech was being described, consider how Sweet (1877) discussed the difficulty in segmenting sounds (in this case “glides”): Consonant-glides are more noticeable in French than in English, especially in stop- combinations, (strik[ʜ]t)=‘strict.’ Final voice stops often end in a voice-glide, (bag[ʌ]) =‘bague.’ In passing from (ɴ) to the next vowel the glide is generally formed so slowly as to be heard as a separate element, so that (ohɴ[i]oq)=‘oignon’ sounds like (ohɴjoq). Final (j) and (ɴ) end voicelessly, the glottis being opened at the moment of removing the tongue from the consonant position, so that (fijʜ) and (vIɴʜ) sound like (fij-jh) and (viʜ- jh). (p. 125) In later attempts to standardize ad hoc signs, alphabet systems came to be used. However the choice of a letter system was circumstantial. In 1886, Passy headed a group of phoneticians and founded the Association phonétique internationale with the purpose of developing a transcription system for teachers of European languages (Auroux, 2000). In 1888, the association published the first version of the International Phonetic Alphabet, which came to be adopted by philologists and phonologists. The latter formally introduced a “commutation test” as a method for establishing the phonological system of a language. With this method, IPA letters are substituted in “minimal pairs” to identify “distinctive features” that differentiate literal meanings in paired lexemes. For instance, in English /bak/-/pak/, the commutation of b-p is seen to create a distinction in meaning in terms of a subjectively estimated “voice” feature. However, letters in such analyses were not merely a descriptive device. They took on a theoretical and psychological 26 Questions of Ontology
- 51. status where features were thought to be actually bundled together at some mental level, in letter-like “phonemes,” if not in signals (see, e.g., early debates on whether phonemes were real elements in signals, “mental images” as Saussure called them, or “convenient fictions”: Twadell, 1935, contra Sapir, 1921, 1933). In considering these developments, the arbitrariness of the standard is not the problem. Many standards in science can be arbitrary to a degree. Rather, it is the choice of culture-specific signs that appears objectionable. There was no attempt to justify the use of alphabet symbols except by reference to their usefulness in teaching European languages (Auroux, 2000). The “International” system of the IPA set aside tone languages and a variety of “suprasegmental” marks that did not fit letters (a problem that was only partly addressed in later versions of the IPA). More fundamentally, there was no theoretical justification for restricting functional distinctions to minimal pairs of isolated words as represented in European-style dictionaries. By this stand- ard, the use of letters in commutation tests overlooked the vast range of distinctions that could convey meaning differences in spoken language. This created problems in defining just what type of meaning phonologists were talking about. In fact, minimal pairs were restricted to distinctions in “literal” meaning (and see Coulmas, 1989). This was a determining point. At a time when the study of spoken languages was beginning to emerge as an academic discipline, the use of letter notations and phonological methods basically limited the object of study to what could be alphabetically transcribed. And then there was the problem of the validity of representing subjectively estab- lished features in successive letters, which prescribed that articulated sound features occur in bundles on a line as “one phoneme per letter” (Jones, 1929). Initially, there was the shared belief amongst philologists and phoneticians that their descriptions of speech sounds were “scientific” and captured actual articula- tory and acoustic units. But following early instrumental observations, some authors expressed doubts. There were problems in deciding how to divide sounds into phonetic descriptions (as exemplified in the previous quote from Sweet). Instrumentalists like Rousselot (1897) and Scripture (1902) saw no basis in the continuous curves of their kymograph recordings for dividing speech beyond respiratory pauses. Both saw that, for the same letter representations, speakers of different languages and dialects (and even different age groups) produced variable sounds such that letter representations appeared illusory. On the other hand, these early works, which contained innumerable descriptions of waveforms, bore few explicit conclusions on speech structure or the validity of language analyses based on alphabetic script. A later critical report by Stetson (1928/1951) addressed classical phoneticians, Saussure, and Prague phonologists more directly. Stetson’s research demonstrated that an analysis of spoken language in terms of letter symbols on a line does not conform to the structural aspects of speech as seen in the “phase” of articulatory motions (i.e., the articulation of a vowel 27 How We Are Introduced to the Study of Spoken Language
- 52. V does not start after a consonant C in producing CV). His report also showed that one could not segment speech into units smaller than syllable-like “ballistic pulses” – that is, one cannot halt speech during a pressure- building constriction of the vocal tract or the vocal folds such that speech minimally involves a cycle of motions that build and release pressure. Stetson explicitly outlined the implications for language analysts by quoting from works in phonology: To isolate and classify the essential “sounds” (phonemes) of a language, i.e. to assemble the phonetic alphabet of a language, was an important project for Sweet and F. de Saussure, and a primary enterprise for Trubetzkoy. This was the impulse that lay behind the International Phonetic Alphabet, the first achievement of the phoneme doctrine. Sweet noted that “sounds” differentiated words; at the hands of Saussure, Trubetzkoy and associates, the smallest phonic change in a word that shifts the meaning was made to indicate a new phoneme. “Dans le ‘projet de terminologie standardisée’ soumis à la Réunion Phonologique Internationale de 1930, on trouve les définitions suivantes: ‘Une opposition pho- nologique est une différence phonique susceptible de servir dans une langue donnée à la différenciation des significations intellectuelles; chaque terme d’une opposition phonologique quelconque est une unité phonologique; le phonème est une unité phonologique non susceptible d’être dissociée en unités phonologiques plus petites et plus simples.’ (Travaux du Cercle Linguistique de Prague, IV, p. 311).” It is apparent that the method involves a resort to meanings and also to observations of articulations. Certainly, in one form or another, differentiating the significant articula- tions by differences of meaning will be important to any system. It is unfortunate that the scholars who have made the most of the method have not only ignored the syllable but have also insisted that the articulation occurs in the separate, concrete world of la parole, while the phoneme symbol occurs in the separate, ideal world of la langue; and so the phonemicists have made a virtue of their ignorance of experimental methods of obser- vation. (p. 136) More importantly, Stetson cautioned language theorists on the writing bias underlying a conceptualization of speech as sequences of letter-like phonemes, and indicated why these units may be impossible in terms of temporal con- straints on speech processing. The series of characters which we read and write as representing an articulate language have given us a mistaken notion of the units which we utter and which we hear. The series of speech units cannot correspond to the series of “sounds” or “phonemes” set down on paper. Even “slow, careful utterance,” let alone the rapid utterance of everyday, is much too fast for that. The maximum rate at which articulations can be uttered is 10–12 per sec.; and the maximum rate at which auditory signals can be identified is 14–16 per sec. A [transcribed] syllable of speech often indicates 2–7 phonemes; “do” has two, “tree” three, “quilt” five, “squelched” seven; the slow, careful rate of utterance is often 4–5 syllables per sec. Thus phonemes are often indicated 15–25 per sec. Obviously the phonemes are not uttered (or heard) one after the other; there must be 28 Questions of Ontology
- 53. extensive overlapping, as physiological tracings prove. The consonants do not prove to be separable units, they must have breath pressure behind them; the pressure is supplied by the pulses of the syllables in which they function. A consonant cannot be pronounced alone; it is always a characterized factor in some syllable. The supposed consonant “elements”, which naive teachers assume that they are uttering, prove to be syllables with pulses from the chest, but often with the vowel shape unvocalized. (p. 137) The idea of syllable-like pulses or cycles is developed later on in this mono- graph, but it is useful to note that Stetson’s wording of “syllables with pulses from the chest” suggested that the pulses link to respiratory mechanisms. Studies using electromyography (EMG) have shown, to the contrary, that expiratory processes are largely passive during normal speech (Draper, Ladefoged, & Whitteridge, 1959; Ladefoged, 1962). According to Ohala (1990, p. 30), Stetson likely misin- terpreted observations of pressure variations in the lungs as indicating an active factor when, in fact, the variations reflect a resistance to expiratory flow created by cyclical articulatory motions upstream (see also Boucher & Lamontagne, 2001). On the other hand, several studies support Stetson’s main claim: speech cannot be segmented in units smaller than syllable-like pulses seen as coherent cycles of pressure-building and releasing motions where the latter can be voiced or unvoiced. Research in the 60s confirmed that C and V motions are not timed as successive phonemes, but “co-articulated” in coherent CV-like units (MacNeilage & DeClerk, 1967, and see Chapter 4). Furthermore, studies combining EMG and photoglottography confirm cyclical motions and pulses as basic units of produc- tion. Thus, sounds that can be produced in isolation, like vowels and “syllabic” consonants, minimally involve a pressure-rising constriction of the vocal folds prior to a voiced release (Hirose, Sawashima, & Yoshioka, 1980, 1983). Stop consonants like [p, t, k, b, d, g], however, cannot be produced in isolation. Articulation of these sounds involves unitary cycles where a pressure-building oral closure entails a releasing motion, and this is observed in CV, post-vocalic contexts as in VC, CVC, and complex clusters where the oral release can be voiceless, or voiced as brief schwas (e.g., Ridouane, 2008). The latter findings, incidentally, suggest that IPA notations and orthographic notions of VC and CVC as single syllables misrepresent the fact that these sequences are produced as two syllable-like cycles of motion each involving separate bursts of EMG activity (Boucher, 2008, and see the demonstrations in Part III). Stetson’s view of unitary cyclical pulses can account for several phenomena such as why certain consonant motions can create syllable beats, or the particular coherence of CV patterns (MacNeilage & DeClerk, 1967), discussed in subsequent chapters. As for the remark on the impossibility of phoneme-by-phoneme processing in production and perception, Stetson’s claim of a maximal rate of about 100 ms/syllable conforms to an observed upper limit on the production of separate sounds (Kent, 1994; Knuijt, Kalf, Van Engelen et al., 2017). In terms of acoustic perception, an extensive review by Warren (2008) indicates that order perception for unstructured 29 How We Are Introduced to the Study of Spoken Language
- 54. speech sounds such as sequences of vowels or digits separated by silences is possible for rates no lower than 100 ms per item. Sequences in which items are artificially compressed to about phoneme-size elements can create auditory illu- sions and confusions in perceived order. If such observations are confirmed for structured speech, they would further substantiate Stetson’s original critique of phonemes as processing units. But findings such as those reported by Stetson were not a serious concern for language theorists of the day. In overlooking instrumental findings, a favored argument of early phonologists that is often repeated in introductory texts, was that instrumentalists could describe speech sounds to no end. Consequently, some saw instrumental observations as having to be guided by considerations of the functional aspects of sounds serving to communicate meaning distinctions. This argument of the “primacy of linguistic criteria” was present in Saussure’s writings, in early phonetics (see Sweet’s notion of “broad” transcriptions), and was essential to the constitution of phonology as a separate subfield of language study. As Trubetzkoy (1939/1971) stated: “It is the task of phonology to study which differences in sound are related to differences in meaning in a given language, in which way the discriminative elements [. . .] are related to each other, and the rules according to which they may be combined into words and sentences” (p. 10). For some phonologists this meant that speech observations had to be hypothesis driven, they had to follow assumptions of linguistic analyses and the criteria of distinctiveness because “nothing in the physical event . . . tells us what is worth measuring and what is not” (Anderson, 1985, p. 41, quoted by Lindblom, 2000; see also Anderson, 1981). It is the case that, to communicate using spoken language, sounds need to be distinctive. But as noted, nothing in the principle of distinctiveness provides an epistemological justification for restricting “features” to sounds that can be transcribed, or to indices that create differences in literal meaning, such as in analyses of minimal pairs of “words” as listed in dictionaries. Nor does this principle serve to justify the view that distinctive features occur in letter-like segments, or any other extent of speech for that matter. The issue of sound segmentation is an empirical question bearing on how features are produced and perceived along a time axis. It is an issue that was not resolved by the choice of alphabet signs to represent speech. On the contrary, in the history of language study, the decision to analyze speech as sequences of letters said to represent phonemes finds no “phonological” or experimental motivation. Yet in choosing to represent sound features with letters, the notion of letter-like packets of features gained a theoretical status, even though this choice was entirely circumstantial. In fact, when the field of language study began to consolidate in the latter part of the nineteenth century, various transcription systems were available. Some systems, like Bell’s Visible Speech, which became popular in some circles, did not involve letters at all (Abercrombie, 30 Questions of Ontology
- 55. Other documents randomly have different content
- 56. officers of the city, and General Sherman occupied one of the proscenium boxes with a party of friends. He seemed to be in the best of health and spirits, and gave every evidence of keen enjoyment of the opera. He returned to his home immediately after the performance, and, although the weather was clear and bright, in some way he caught a severe cold. Its first effects were noticed on the following morning. His condition, however, did not prevent his attendance at the wedding of Miss Shepard, daughter of Colonel Elliott F. Shepard, on that afternoon. He coughed a little and complained of the cold while in the church. On Friday morning his condition had become more uncomfortable, but excited no alarm. His throat, however, had become affected in the meantime, and he was obliged to give up a dinner with Lawrence Barrett that evening at the Union League Club. On Saturday morning when he began to show signs of facial erysipelas, accompanied by fever, he felt some anxiety, and sent for Dr. C. J. Alexander, a surgeon of the army, who had been his family physician for a number of years. On Sunday the disease began to get a firm hold upon the old warrior. His face and neck became much swollen and inflamed, and conversation became difficult and painful. His condition was such that Dr. Alexander sent for Dr. Janeway, for the purpose of holding a consultation. The General was then confined to his bed, and it was found that the ordinary treatment applied in cases of erysipelas would not answer the purpose, in part owing to the General's advanced age. Sunday, by the way, was the seventieth anniversary of his birth. The disease had developed to such an extent on Monday that it was decided to summon the members of the family. Telegrams were sent at once to Senator John Sherman, his brother; his daughters, Mrs. Thackara and Mrs. Fitch. The other children, with the exception of the Rev. T. E. Sherman, were at home. To him, however, a cable dispatch was sent. He was a student in the Jesuit Seminary on the Island of Jersey. Senator Sherman arrived at his brother's home on Monday night, and his daughters on the following day. The arrival of Senator Sherman, with the publication of the dispatch which called
- 57. him, was the first intimation that the people of New York City had of General Sherman's illness. Dr. Alexander remained at the sick man's bedside on Tuesday night, and when Dr. Janeway came to relieve him on Wednesday morning, February 11, he found the General resting on his back in a state of semi-stupor. His condition at that time was recognized as critical. He was in great pain when he moved, and gave evidence of growing weaker, despite the fact that whiskey and milk, which were used as nourishment throughout the illness, were administered to him as often as possible. Intimate friends of the family were then informed of his precarious condition. The General rallied somewhat at noon, and his family began to hope that the illness was only temporary. But their hopes were delusive. In the afternoon, the attending physicians, Drs. Alexander, Janeway and Greene, began to send out hourly bulletins as an official answer to the hundreds of inquiries that poured in upon them. At 2.15 they made their first announcement, which read as follows: "General Sherman was worse this morning, and his condition is critical. During the day his condition has improved considerably." About 5 P.M. General Ewing said that he had called on General Sherman, and had been recognized by him. As soon as he saw General Ewing enter the room, the patient called out, "Hello, Ewing." He did not make any attempt to sustain conversation, however. His enunciation was difficult, and, besides, though he could recognize his friends, he did not seem to be able to have enough energy or command of his faculties to talk to them. He improved again slightly during the evening, so that two of the physicians and Senator Sherman left the house. The Senator, however, was recalled at two o'clock on Thursday, when the veteran again grew worse. Thursday passed in much the same way as Wednesday, although it was deemed advisable by the family, for their comfort, to have the last rites of the Catholic Church administered to him, just before noon. In the afternoon the sick man surprised his watchers by getting out of bed and walking a few steps
- 58. to an easy chair, where he sat for a few moments. He showed the same marvellous will power again in the evening. In his rallies he was able to clear his lungs a little. Whiskey and milk were given to him as often as he could take nourishment. Late at night it was said that if the General could maintain his state till that time there would be hopes of ultimate recovery. Friday was another day of hope and disappointment. Several times it was reported that the General was dying, but he managed to rally despite his weakened condition. Said General Ewing that evening: "Sherman is perfectly conscious, and when spoken to rouses up and makes a perfectly intelligible answer to any question that may be asked. He is deaf, you know, and it is necessary to address him in a pretty loud voice, in order to be heard." "Does he recognize his friends?" "Not until spoken to, and I doubt if he recognizes them even then. I doubt if he has recognized me in the last two days." "Yet he talks to them?" "He does not talk much. The tongue is much swollen and the jaw is stiff, and he can speak only with difficulty." "Does he realize the serious character of the disease?" "It is hard to say. He has given no evidence of uneasiness, except when he called for 'Cump' (P. T. Sherman, his son), on Thursday. It then occurred to me that he wanted to say a last word to the young man. But I may have been mistaken. At any rate, when 'Cump' went to him he was unable to tell him what was on his mind." The illustrious patient grew weak again at midnight, and at an early hour Saturday morning, February 14, it was known that his death was only a question of a few hours. At four o'clock his family was all summoned to his room and never left it, except for a few minutes, until the end. The alarming attack which seized the patient soon after six o'clock precipitated death. The doctors hurriedly held
- 59. another consultation, did what they could to relieve his distress and then decided that hope must be abandoned. The chloroform plasters which had been placed on Sherman's chest, failed to help. The police officers then cleared the sidewalk and streets of all passengers, and people began to wait for the end. At 8.35 o'clock Dr. Janeway left the house, to which he did not again return. His face and his few words told plainly that he had no hope. About half an hour before the General's death the watchers discerned signs of approaching dissolution. First the old soldier's fingers began to grow cold, then the fatal coldness crept slowly up his arms, and over his body. As the end approached, the General's head, which had been resting on a large pillow, was lowered gradually in the hope that he might be enabled to breathe easier. Although he died from suffocation, caused by the mucus from his inflamed throat filling his lungs, there were no longer indications of suffering on his part. Those who were nearest his head say that they heard a gentle sigh escape his lips and then all was over. It was just 1.50 o'clock when the famous soldier expired. There was no clergyman of any denomination in the house during the day. Within a minute or two after General Sherman's death one of his men-servants stepped outside of the front door and said: "It is all over." Kneeling at the bedside, as the soldier's spirit left its earthly tenement, were the General's son, P. T. Sherman, his four daughters, the Misses Rachel and Lizzie Sherman, Mrs. Fitch and Mrs Thackara; his brother, Senator John Sherman; his sons-in-law, Lieutenants Fitch and Thackara; his brother-in-law, General Thomas Ewing; his physician, Dr. Alexander, U. S. A., and his nurse, Miss Elizabeth Price, of the New York Hospital. The other son, the Rev. Thomas E. Sherman, was on the ocean, hastening homeward, but too late. Generals Slocum and Howard were then in the room below. General Sherman seems to have had a presentiment of his fate some weeks before it actually befell him. One day he said to General
- 60. C. H. T. Collis, who mentioned Grant's birthday—April 27: "Oh, well, Collis, I'll be dead and buried before then." "I tried hard to cheer him," said General Collis, "and pretended to believe he was joking, but he became serious and added after awhile: 'I feel it coming sometimes when I get home from an entertainment or banquet, especially these winter nights. I feel death reaching out for me, as it were. I suppose I'll take cold some night and go to bed, never to get up again.' The words were prophetic." In accordance with General Sherman's often expressed desire, the body did not lie in state; and the public so respected the grief of the family as not to besiege the house to gaze upon the remains of the hero. General Howard sent over a guard from the army post on Governor's Island, and with General Slocum, by invitation of the family, took charge of the arrangements for the funeral obsequies. The body of the deceased General was placed in a coffin exactly like that in which Mrs. Sherman was buried. The General chose her's himself, and gave express orders that his own should be like it. It was of oak, lined with cream-colored satin, and had silver handles. On a silver plate was the following inscription: WILLIAM TECUMSEH SHERMAN, GENERAL, U. S. A. Born February 8, 1820. Died February 14, 1891. This coffin was inclosed for the journey to St. Louis in an outer coffin of chestnut wood, brass bound, with a brass plate bearing the same inscription as the inner. The General's body was dressed in the full uniform of his rank. The following "Special Order No. 5" was issued from the headquarters of the Grand Army of the Republic, at Rutland, Vt. "Grand Army of the Republic posts on the route of the funeral train of General Sherman from New York to St. Louis will form at
- 61. their respective railroad stations and salute remains as train passes." The President and his Cabinet were invited by General Howard to attend the funeral exercises in New York. Committees from both Houses of Congress were appointed to pay their tribute of respect. From the Senate came Messrs. Evarts, Hawley, Manderson, Pierce, Cockrell and Walthall. From the House Speaker Reed appointed Messrs. Cutcheon, Spinola, Cogswell, Cummings, Grosvenor, Kinsey, Tarsney, Henderson, of Illinois, and Outhwaite. A sorrowful meeting of the Military Order of the Loyal Legion was held on Monday, February 16, at which these resolutions were adopted: "In common with the entire country we lament the loss of a great military chieftain whose loyal spirit rightly placed the love of country higher than all earthly obligations, and who was individually a distinct and glorious element in the triumphant struggle of that country for its own survival and for the rights of man. "As once his fellow soldiers we mourn universally for the dead commander, whose great heart made us all his own and made his own virtues seem to us like personal benefactions. "As members of this Military Order we deplore the loss of a companion whose honors added to the value of those ties which his fellowship helped to endear, and whose frequent and cordial visits to the New York Commandery will be cherished in our memories as so many occasions to be often and affectionately recalled. "To his children and relatives, to whom his great renown, his honors and his tenderness do but enhance their loss, we tender all that sympathy may, and trust that a place in our regard henceforth may be accepted by them as a little heritage from him." General Howard made a brief address, in the course of which his emotion was strong and interrupted his utterances. "General Sherman," he said, "had more personal friends and could call more men by name probably than any other man in the
- 62. country. "A few days ago, Sherman and Slocum and I met in Brooklyn and the conversation turned on death. Some one remarked that he hoped it would not come to Sherman for many years. I exclaimed, on the impulse of the moment, 'General, you will never die.' He answered, sharply and strongly, 'My body will die.' God bless General Sherman," was the peroration of General Howard's speech. General Slocum followed with a warm panegyric on the march to the sea. "Sherman was to me something more than a companion," he said. "He gave me his confidence in war and his friendship in peace. He opened to me what is dear to every soldier, an opportunity to link my name with his. "In the coming time there will be no dispute about his career. It may be in the future that some man will say that he furnished the idea of the march to the sea to Sherman. That man must have been with him at the time, or subsequent, when Sherman captured Atlanta, for when he did so he had no idea of cutting aloof from his base of supplies. When he got back from the battle of Jonesboro he took down a map and said, 'I will make Atlanta my base of supplies.' He went so far as to throw up intrenchments. That was before Hood pushed up toward the Tennessee and Nashville; and then he changed his mind. "After Sherman had taken Savannah certain persons at Washington urged him to take his troops to City Point by sea. Had he been a timid man he would have been content to rest upon his laurels, knowing that he had already won an imperishable fame, but he said: 'No; I will take my chances in South Carolina,' and he did so, and everything went like clockwork, and success again crowned his efforts." At the same time a meeting of representative citizens of St. Louis was held in that city to make arrangements for the final services there; and every city and town along the route prepared to salute the funeral train with demonstrations of sympathy and honor.
- 63. The orders for the procession in New York were issued on February 18, as follows: Headquarters of the Atlantic, Governor's Island, New York. The arrangements for the funeral of the late illustrious General of the Army, William Tecumseh Sherman, having been entrusted by his children and other relatives to the care of the undersigned, they have agreed upon the details so far as they relate to the ceremony in New York, which are now furnished for the information and guidance of all who may participate therein: The regulation escort, under command of Loomis L. Langdon, 1st Artillery, will consist of one regiment of United States marines, four companies of United States engineers, and six companies foot batteries of artillery; of a battalion of light artillery from the Army and the National Guard of New York, and of two troops of cavalry from the National Guard of New York. The remains will be received by the escort at the late residence of the General, No. 75 West Seventy-first street, at 2 o'clock, P. M., on Thursday next, the 19th inst. The body will be borne on a caisson, preceded by the following-named pall- bearers in carriages: Major-General J. M. Schofield, Major- General O. O. Howard, Rear-Admiral D. L. Braine, Rear- Admiral J. A. Greer, Professor H. L. Kendrick, Major-General H. W. Slocum, General Joseph E. Johnston, Major-General D. E. Sickles, Major-General G. M. Dodge, Major-General J. M. Corse, Major-General Wager Swayne, Major-General Stewart L. Woodford, Brigadier-General Jno. Moore, Brigadier-General H. G. Wright. These pall-bearers will accompany the remains as far as the train at Jersey City. Six sergeants will proceed to St. Louis. The special escort of honor from the Grand Army, Lafayette Post, will form on the right and left of the caisson.
- 64. The order of column following the family and relatives will be as follows: (1) The President and Vice-President of the United States. (2) The members of the Cabinet. (3) Ex-Presidents of the United States. (4) Committees of the Senate and House of Representatives. (5) The Governor of the State and the Mayor of the City of New York. These officers will follow the family and relatives as representative mourners. (6) The Military Order of the Loyal Legion of the United States, and officers of the Army and Navy. (7) The Grand Army of the Republic. (8) The Corps of Cadets, United States Military Academy, Lieutenant-Colonel Hawkins commanding. (9) The National Guard, under Brigadier-General Louis Fitzgerald. Delegates and representatives from veterans, sons of veterans and other organizations unassigned, under charge of General David Morrison. The line of march will be as follows: Eighth avenue to Fifty-ninth street, to Fifth avenue, to Broadway, to Fifty- seventh street, to Fifth avenue, to Washington Square: there the column, excepting the regulation military escort, will be dismissed. This escort will continue its march by Waverley Place to Macdougal street, to King street, to Hudson street, to Watts street, at corner of Canal, through Watts street to junction with West street.
- 65. Veteran organizations not moving with column will form across West st. from Watts st. to the ferry landing, foot of Desbrosses st. The carriages in the procession will be restricted to the pall-bearers, family and relatives, and invited guests. The column will be commanded by Major-General O. O. Howard, United States Army. Major-General Daniel Butterfield is designated as senior aide to the General Commanding and as marshal. The following aides are announced: General Horace Porter, to accompany the President of the United States; General M. D. Leggett, to accompany the Cabinet; the Hon. Joseph H. Choate, to accompany ex-President Hayes; the Hon. Chauncey M. Depew, to accompany ex-President Cleveland; General Floyd Clarkson, in charge of the Grand Army; Major-General H. A. Barnum, to accompany the Superintendent of the Military Academy; General Robert Nugent, formerly of General Sherman's regiment, to take charge of the veterans at Desbrosses st. David Morrison, 79th Veterans, in charge of veteran organizations in columns other than the Grand Army; Mr. Kiliaen Van Rensselaer, to accompany carriages of relatives. Mr. Loyall Farragut. Captain H. P. Kingsbury, 6th Cavalry. Captain A. M. Wetherill, 6th Infantry. First Lieutenant L. A. Craig, 6th Cavalry. First Lieutenant Guy Howard, 12th Infantry, Aide-de- Camp. First Lieutenant Harry C. Benson, 4th Cavalry. First Lieutenant Charles G. Treat, 5th Artillery, Aide-de- Camp.
- 66. First Lieutenant W. W. Forsyth, 6th Cavalry; Second Lieutenant Samuel Rodman, 1st Artillery, Aides-de-Camp. The churches of New York City are requested to have their bells tolled at half-minute intervals during the movement of the columns, from 2 until 4 P. M.; and the churches of Jersey City are requested to toll their bells in like manner from 5 to 6 P. M., on Thursday. The headquarters of the General commanding the column and the Marshal, will be announced to-morrow. The details of the formation in line of the respective divisions will be communicated to the commander or chiefs from headquarters. H. W. Slocum. Oliver O. Howard. Late on Wednesday night the steamship Majestic arrived at New York, with the Rev. Thomas E. Sherman among its passengers. When the pilot boarded her, Mr. Sherman eagerly asked him about the General. "I'm unable to say," replied the pilot, adding that, he only knew of General Sherman's sickness, as he had been out at sea for some days. When the mail steamer came alongside, Mr. Sherman repeated his anxious inquiry. The answer came back, "General Sherman's funeral takes place to-morrow."
- 67. ARMY AND CORP COMMANDERS OF THE ARMY OF THE TENNESSEE. Sherman. Logan. Grant. Dodge. Blair. McPherson. Howard. From painting by J. E. Taylor. The day before the funeral the house was opened for a few hours, and the public were allowed to enter and view the face of the dead. Thousands availed themselves of the privilege. "It was an interesting crowd of people. There were white-haired veterans of the war; there were people in the clothing of luxury, people clad like beggars, and mothers with babies in their arms leading children by the hand. There were schoolboys come to look at the man about whom their histories tell them, come to see if the face they had seen in the pictures was indeed the face of the great General. There were young girls there, and young men also. It was a crowd representative of the whole American people. Hebrews came out of the depths of the east side and Germans came from Hoboken. All
- 68. passed in review before the man who will review armies no more. Their uncovered heads were bowed. Some of the very old women who had given their sons to this leader for their country's sake sobbed as they passed on." It was on a glorious winter day, February 19, that the dust of the great soldier was carried from his former home to make the journey to its final resting place at St. Louis. As the funeral procession started, bells of the City were tolled; buildings everywhere displayed tokens of honor and signs of mourning; the streets were thronged with sympathetic spectators; and thirty thousand men marched with measured tread behind the coffin that contained the earthly remains of their loved and honored leader. Conspicuous in the company were General Schofield, the head of the army; General Howard and General Slocum, Sherman's lieutenants on the march through Georgia; General Corse, of Kenesaw fame; General Johnston, Sherman's old antagonist; and Professor Kendrick, one of those who taught Sherman the art of war. The President, the Vice-President, the two living ex-Presidents, and the members of the Cabinet were also in the company. There was a large contingent from the regular army, with General Howard in command. Then came the Military Order of the Loyal Legion; long columns of the Grand Army of the Republic; West Point Cadets; the Sons of Veterans; and delegations from various clubs, commercial organizations, and the municipal government. The long procession wound its way through the streets of New York to the Jersey City ferry. There the coffin and its immediate escort were taken across the river and placed on the funeral train. General Sherman's horse, which with empty saddle had followed the funeral caisson, was led up to the train and the saddle and boots were placed by the coffin in the funeral car. The train consisted of an engine and eight cars. Generals Howard and Slocum, and Surgeon Alexander, besides six sergeants of the regular army, acted as a guard of honor. The Governor of New Jersey through his staff acted
- 69. as an escort through Jersey City; and the Governor of Pennsylvania and his staff in a special car went through to Harrisburg. It was early in the evening when the train left Jersey City. At almost every station that it passed vast throngs assembled and bands of music played solemn dirges. It was midnight when it reached Harrisburg, Pennsylvania, yet a multitude stood in the darkness in the open air to do it honor. In the morning it passed through Pittsburg in the midst of a heavy rain storm. Later in the day the sky was clear and the sun shone brilliantly. At Steubenville, Ohio, seventy-five veterans of the army stood on the platform as the train went by, nearly all of them old comrades of Sherman. At Columbus, Ohio, the train paused for a few minutes while Grand Army veterans were allowed to gaze upon the casket. At Indianapolis another stop was made while many distinguished people paid their tribute of honor to the mighty dead. It was Saturday morning when the train reached St. Louis. For several days the weather there had been stormy, but this morning the skies were clear and the sunshine bright. Thousands of people thronged about the station, waiting there for hours before the arrival of the train. At last, at a little before nine o'clock, the funeral cars slowly rolled into the station, the engine bell solemnly tolling. Elaborate preparations had been made at St. Louis for a military funeral befitting the great soldier whose dust was to be returned to the dust from which it came. Two hours after the arrival of the train the procession was formed, under the lead of General Wesley Merritt, and it solemnly wound its way through the city which for many years was Sherman's favorite home, to Calvary Cemetery. The first division consisted of detachments of the Regular Army, escorting the casket, which was borne on a caisson drawn by four black horses and covered with the stars and stripes. Ransom Post, No. 131, Department of Missouri, G. A. R., acted as the immediate guard of honor. Following closely were the members of the President's Cabinet and the committees from the two houses of Congress. The second division was made up of the Loyal Legion and
- 70. the Society of the Army of the Tennessee. In it were ex-President Hayes, Judge Gresham and General Lew Wallace. The third division was composed of Posts of the Grand Army of the Republic and Sons of Veterans. In the fourth division were militia regiments from various States and many civil officials. Civic societies made up the fifth division, and various city delegations and the general public the sixth and last division. As the long procession wound its way up the slope to Calvary Cemetery it presented a view of solemn but inspiring splendor. The arms of the troops flashed in the sun-light, a multitude of flags fluttered on the breeze, and the subdued strains of funeral music made the air tremulous. At last, six miles from the railroad station, the plot was reached where were the graves of the wife and two children of the departed hero. The flag covered casket was borne upon the shoulders of eight sturdy soldiers to the open grave. Then came the command, "Present Arms!" And every soldier stood motionless as a graven statue. Then the Rev. Thomas E. Sherman, clad in slight vestments, stepped forward and began the service for the dead over his father's dust, standing, as he did so, in the shadow of his mother's monument. He repeated the words of the Litany, translating prayer and scripture into English, in a clear, manly voice, and offered a touching extemporaneous prayer. After the last solemn words a company of troops stepped forward. Three times were given the commands, "Load!" "Ready!" "Aim!" "Fire!" and three times the rifles spoke their loud farewell salute. Then the artillery posted near by thundered forth their echoing responses. When the last reverberations died away a solitary trumpeter stepped forward to the foot of the soldier's grave and sounded "Taps." Thus ended the last impressive scene. In his life Sherman had left with his friends full instructions concerning his funeral, his grave and his monument. He directed that the only inscription above his dust should be his name, his rank, the date of his birth, the date of his death, and the simple words,
- 71. "True and Honest." A fitting epitaph for one who was truly, as was written of another great soldier, "In his simplicity sublime."
- 72. CHAPTER XXX. TRIBUTES. A National Outburst of Grief—The President's Message to Congress— The Senate's Memorial Resolutions—Senator Hawley's Eulogy—A Touching Tribute from a Southern Senator—Speeches by Senators who were also Soldiers—Eloquent Words from Lawrence Barrett— Judge Gresham Recalls Sherman's Prophetic Words—A Comparison Between Sherman and Lee—General Slocum's Reminiscences— Chauncey Depew on Sherman in Social Life. During General Sherman's last illness the entire nation listened with anxious suspense to every word of news that came from his home, and millions of hearts hourly offered fervent prayers for his recovery. The announcement of his death was not unexpected, for it had been known for several days that recovery was impossible; but it was none the less a shock to the public. Everywhere expressions of grief were heard and emblems of mourning were seen. Flags were placed at half-mast and buildings draped in black; bells were tolled and memorial meetings held. Messages of sympathy and condolence came to his family by mail and telegraph from every part of the world. Only a few irreconcilable spirits here and there in the South spoke against him, and made his death an occasion for venting their spleen against the patriot who had subdued the rebellion. When the news of Sherman's death reached Washington, the President, who had himself been an officer in Sherman's army in
- 73. Georgia, sent a message announcing the fact to Congress, in which he said: "The death of William Tecumseh Sherman is an event that will bring sorrow to the heart of every patriotic citizen. No living American was so loved and venerated as he. To look upon his face, to hear his name, was to have one's love of country intensified. He served his country, not for fame, not out of a sense of professional duty, but for love of the flag and of the beneficent civil institutions of which it was the emblem. He was an ideal soldier, and shared to the fullest the esprit de corps of the army; but he cherished the civil institutions organized under the Constitution, and was a soldier only that these might be perpetuated in undiminished usefulness and honor. He was in nothing an imitator. "A profound student of military science and precedent, he drew from them principles and suggestions, and so adapted them to novel conditions that his campaigns will continue to be the profitable study of the military profession throughout the world. His general nature made him comrade to every soldier of the great Union Army. No presence was so welcome and inspiring at the camp-fire or commandery as his. His career was complete; his honors were full. He had received from the Government the highest rank known to our military establishment, and from the people unstinted gratitude and love. No word of mine can add to his fame. His death has followed in startling quickness that of the Admiral of the Navy; and it is a sad and notable incident that when the Department under which he served shall have put on the usual emblems of mourning, four of the eight Executive Departments will be simultaneously draped in black, and one other has but to-day removed the crape from its walls." Senator Hawley, of Connecticut, at once offered the following resolutions, which were unanimously adopted by the Senate: "Resolved, That the Senate receive with profound sorrow the announcement of the death of William Tecumseh Sherman, late General of the armies of the United States.
- 74. "Resolved, That the Senate renews its acknowledgment of the inestimable services which he rendered to his country in the day of its extreme peril, laments the great loss which the country has sustained, and deeply sympathizes with his family in its bereavement. "Resolved, That a copy of these resolutions be forwarded to the family of the deceased." Mr. Hawley said: "Mr. President, at this hour, the Senate, the Congress and the people of the United States are one family. What we have been daily expecting has happened; General Sherman has received and obeyed his last order. He was a great soldier by the judgment of the great soldiers of the world. In time of peace he had been a great citizen, glowing and abounding with love of country and of all humanity. His glorious soul appeared in every look, gesture and word. The history of our country is rich in soldiers who have set examples of simple soldierly obedience to the civil law and of self- abnegation. Washington, Grant, Sheridan and Sherman lead the list. Sherman was the last of the illustrious trio who were by universal consent the foremost figures in the armies of the Union in the late war. Among the precious traditions to pass into our history for the admiration of the old and the instruction of the young was their friendship, their most harmonious co-operation, without a shadow of ambition or pride. When General Grant was called to Washington to take command of the armies of the Union, his great heart did not forget the men who stood by him." Here Mr. Hawley read the letter from Grant to Sherman, written at that time, expressing thanks to him and McPherson as the men, above all others, to whom he owed success, and Sherman's letter, in reply, saying that General Grant did himself injustice and them too much honor. Mr. Hawley closed his remarks, his voice frequently giving way from grief and emotion, by reading the following passages from Bunyan's "Pilgrim's Progress": "After this it was noised about that Mr. Valiant-for-Truth was taken with a summons. When he understood it
- 75. he called for his friends and told them of it. Then said he, 'I am going to my fathers; and though with great difficulty I got hither, yet now I do not repent me of all the trouble I have been at to arrive where I am. My sword I give to him that shall succeed me in my pilgrimage, and my courage and skill to him that can get them. My marks and scars I carry with me, to be a witness for me that I have fought His battles who will now be a rewarder.' When the day that he must go hence was come many accompanied him to the river side, into which as he went he said: 'Death, where is thy sting?' And as he went down deeper he said: 'Grave, where is thy victory?' So he passed over and all the trumpets sounded for him on the other side." Senator Morgan, of Alabama, said: "On this occasion of National solemnity I would lead the thoughts and sympathies of the American Senate back to those days in our history when General Sherman was, by a choice greatly honorable to his nature, a citizen of the State of Louisiana, and presided over a college for the instructions of Southern youth in the arts of war and the arts of peace. Those were not worse days than some we have seen during the last half of this century. In those days, notwithstanding the conditions of the South, in view of its institutions inherited from the older States of the East, every American was as welcome in Louisiana and the South as he was elsewhere in the Union. We are gradually and surely returning to that cordial state of feeling which was unhappily interrupted by the Civil War. "Our fathers taught us that it was the highest patriotism to defend the Constitution of the country. But they had left within its body guarantees of an institution that the will of the majority finally determined should no longer exist and which put the conscience of the people to the severest test. Looking back now to the beginning of this century and to the conflict of opinion and of material interests engendered by those guarantees, we can see that they never could have been stricken out of the organic law except by a conflict of arms. The conflict came, as it was bound to come, and Americans became enemies, as they were bound to be, in the settlement of
- 76. issues that involved so much of money, such radical political results and the pride of a great and illustrious race of people. The power rested with the victors at the close of the conflict, but not all the honors of the desperate warfare. Indeed, the survivors are now winning honors, enriched with justice and magnanimity, not less worthy than those who won the battles in their labors to restore the country to its former feeling of fraternal regard and to unity of sentiment and action and to promote its welfare. The fidelity of the great General who has just departed in the ripeness of age, and with a history marked by devotion to his flag, was the true and simple faith of an American to his convictions of duty. "We differed with him and contested campaigns and battlefields with him; but we welcome the history of the great soldier as the proud inheritance of our country. We do this as cordially and as sincerely as we gave him welcome in the South, as one of our people, when our sons were confided to his care, in a relation that (next to paternity) had its influence upon the young men of the country. The great military leaders on both sides of our Civil War are rapidly marching across the border to a land where history and truth and justice must decide upon every man's career. When they meet there, they will be happy to find that the honor of human actions is not always measured by their wisdom but by the motives in which they had their origin. I cherish the proud belief that the heroes of the Civil War will find that, measured by this standard, none of them on either side were delinquent, and they will be happy in an association that will never end—and will never be disturbed by an evil thought, jealousy or distrust. When a line so narrow divides us from those high courts in which our actions are to be judged by their motives, and when so many millions now living, and increasing millions to follow, are to be affected by the wisdom of our enactments, we will do well to give up this day to reflection upon our duties and (in sympathy with this great country) to dedicate the day to his memory. In such a retrospect we shall find an admonition that an American Senate should meet, on this side of the fatal line of death, as the American Generals meet on the other side, to render
- 77. justice to each other and to make our beloved country as happy, comparatively, as we should wish the great beyond to be to those great spirits." Senator Manderson said that as the hours of the last two or three days passed away he had not had the heart to make such preparation for the event which he had feared and dreaded, as might seem to be meet and appropriate. The death of General Sherman came (although one might have been prepared for it) as the unexpected. It was a day of mourning and grief. Here, at the Capital of the Nation, lay the body of the great Admiral, the chief of the Navy; and in New York was being prepared for the last sad rites the corpse of the greatest military genius which the Nation had produced. General Sherman had been great not only as a military leader, but he had been great as a civilian. Who was there that had heard him tell of the events of his wonderful career who had not been filled with admiration and respect for his abilities? It seemed to him that General Sherman was perhaps the only man in the North who, in the early days of the war, seemed to appreciate what the terrible conflict meant It was recollected how it was said in 1861 that he must be insane to make the suggestions which he made. These suggestions were so startling to the country that he (Mr. Manderson) did not wonder that men doubted General Sherman's sanity. Like men of great genius, he seemed to have lived in that debatable ground existing between the line of perfect sanity and insanity'. After a review of General Sherman's military career, opening at Shiloh and closing at Atlanta, Mr. Manderson read General Sherman's letter to the Mayor and Common Council of Atlanta, beginning: "We must have peace, not only at Atlanta, but in all America." In conclusion. Mr. Manderson said: "General Sherman was estimable as a citizen, and as fully appreciated the duties of a civilian, as he was admirable as a soldier. But this strife, which we have watched for the last few days, has ceased. The conflict has ended. The Nation has witnessed it. Sixty millions of people have stood in silence, watching for the supreme result. Death, ever
- 78. victorious, is again a victor. A great conqueror is himself conquered. Our Captain lies dead. The pale lip sayeth to the sunken eye: 'Where is thy kindly glance? And where thy winning smile?'" Senator Davis said he could hardly trust himself to speak. He had been a soldier under General Sherman, and had received acts of kindness from him when he was a subaltern. As the years had gone by, and the widening avenues of life had opened up ways of promotion, that acquaintance had ripened into friendship, and, he might say, into intimacy. He had first seen General Sherman at the siege of Vicksburg, twenty-eight years ago, when he was the very incarnation of war; but to-day that spirit had taken up its rest in the everlasting tabernacle of death. It was fit that the clanging of the great city should be hushed in silence, and that the functions of government should be suspended while the soul of the great commander was passing to Him who gives and Him who takes away. No more were heard the thunders of the captains, and the shouting. The soul of the great warrior had passed and was standing in judgment before Him who was the God of Battles, and was also the God of Love. Senator Pierce, as one of the soldiers who had served under General Sherman in the Army of the Tennessee, gave some reminiscences of the war and paid a glowing eulogy to his old commander. Senator Evarts said that the afflicting intelligence of the death of General Sherman had touched the Senate with the deepest sensibilities; that that grief was not a private grief; nor was it limited by any narrower bounds than those of the whole country. The affections of the people toward its honorable and honored men did not always find a warm effusion, because circumstances might not have brought the personal career, the personal traits, the personal affectionate disposition of great men, to the close and general observation of the people at large. But of General Sherman no such observation could be truly made. Whatever of affection and of grief Senators might feel was felt, perhaps, more intensely in the hearts
- 79. of the whole people. To observers of his death, as they had been of his life, General Sherman had been yesterday the most celebrated living American. He was now added to that longer and more illustrious list of celebrated men of the country for the hundred years of National life. One star differed from another star in glory, but yet all of those stars had a glory to which nothing could be added by eulogy, and from which nothing could be taken away by detraction. They shone in their own effulgence, and borrowed no light from honor or respect. It had been said already that General Sherman was the last of the commanders. If those who had passed out of life still watched over and took interest in what transpired in this world (and no one doubted it), what great shades must have surrounded the death-bed of General Sherman! And who could imagine a greater death-bed for a great life than that which had been watched over in a neighboring city during the week? It had been reserved for him (Mr. Evarts) at the declining hour of the day, as a Senator from the State which General Sherman had honored by his late home, and in which he had died, to move, out of respect to his memory, that the Senate do now adjourn. Lawrence Barrett, the eminent actor, paid this eloquent tribute to his friend in the columns of The New York Tribune: "The funeral cortege has passed. The emblems of war, which had for many years been laid aside, have once again been seen sadly embellishing the soldierly equipage whereupon the lifeless body rests. Old comrades, lifelong friends, statesmen and great civilians have followed the mournful pageant with fruitless regrets. The instruments which in battle days sounded to the charge or the retreat, which sang reveille to the waking morn or gave the sternest good-night, when all was well; which through a quarter of a century of peace have greeted the retired warrior at feast and civic parade with harmonies upon his achievements—these now beat the last mournful cadences leading to an earthly camping-ground beneath whose sod the mortal remains of our great soldier shall rest beside his loved ones, forever dead to triumph or threnody.
- 80. "The last of the immortal trio has joined his waiting comrades. Already in the fields of the blessed one may believe that their spirits sadly regard our simple tributes to the earthly casket which holds the dust of Sherman. The mourning thousands who have lined the highway of the sad procession have gone to their homes with a tenderer reflection upon the meanings of existence and death. And even as his valor in the written story had awakened a stronger patriotism than had before existed, so in his death and in the last tributes paid to the hero a fresher and purer sense of patriotic duty springs up in our hearts to link us to the inheritance he helped to gain. "History will gather up and weave into enduring form the achievements of the soldier and the statesman. In that final summary sectional prejudice and personal bias may bear their natural parts. Only in a remote future, when all the sorrowful effects of the great Civil War have lost their nearness—only when its beneficence in knitting closer the bonds of friendship and National brotherhood shall be recognized, when no newly-made grave sends up reproachful reminders to bereaved hearts, only then can the hero's place be immutably fixed on the heroic calendar. To the scholar and the sage may be left that office. The records of his military life, his general orders, his plans, his deeds, will guide the historian into a proper estimate of the dead soldier's station in the military Valhalla. "But how shall the innumerable civic deeds of this dead man be recorded or find place for reference? In the musty archives of no war office are they registered. Upon no enduring parchment are they written. They would escape definition in the attempt to define them. They are engraved upon hearts still living—they sweeten the lives still unsummoned—they are too sacred for utterance. Yet they are the crown of Sherman's achievement. Wherever this man's hand was extended it brought glad strength; wherever his voice was heard it aroused emotions of grateful tenderness; wherever his form was seen it gladdened loving eyes. He survived a civil war for a quarter of a century—to show to us that the soldier's armor is less becoming
- 81. than the garb of civil life, that the pomp and circumstance of war are loud preludes of beneficent peace. "No intrusion of personal relation shall sully this poor testament to the dead. No one can claim the inheritance of such a large- hearted bounty. But in the name of the drama which he loved, in the names of the actors whom he respected, it is proper that no tardy recognition should follow his death. He had a scholar's love for what was highest in the art—whether in the walk of tragedy or comedy. He had a warm affection for those who labored in this atmosphere. He had also a large sympathy for those performances which afford recreation and amusement to the largest class of the community. His voice was never hushed when called to aid in the needs of the player. He was no ordinary first-nighter. He had a simple and affecting belief that his presence might be useful to those who were seeking public suffrage across the foot-lights, and he could not but know that his indorsement was valuable and trustworthy. He was one of the incorporators of 'The Players,' upon whose muster-roll no nobler name appears. His imposing character gave dignity to those deliberative meetings out of which that organization grew into its present useful life. "And should contemporary history fail to do him justice—should the bitterness of the Civil War make a just estimate of his worth impossible in biographical annals—should envy or malice deface the white shaft which should symbolize his deeds—then the dramatist will lovingly bear up the garments of his glory—keep them from soil within that Valhalla where Cæsar and Alexander, Frederick and Gustavus, live imperishably enshrined. Therein shall be cherished the insignia and the characteristics of the most notable figure of modern or ancient soldiery. "Again in future nights shall we see the pomp and glory of Union making war—once again its gallant leader shall pass before the eyes of a curious posterity in the drama's immortal keeping, and the gallant spirit whose influence in life so often attended the presentment of Cæsar and Antony and Cassius and the Roman
- 82. group shall, in death, mingle with their essence, tenderly restored by the dramatists whom he inspired, by the actors whom he loved." HEAD OF PROCESSION COMING DOWN BROADWAY, NEW YORK CITY. Said Walter Q. Gresham, United States Judge: "I belonged to General Sherman's command when he entered Kentucky, at Louisville, in the summer of '61, since which time we have maintained an unbroken friendship. "Besides being a man of great genius he was generous, frank and confiding. No officer of high rank whom I met during the war was more patient than General Sherman with subordinates, so long as he believed that they were trying to do their duty; and no officer was more merciless in dealing with shirks, cowards and pretenders.
- 83. "In brilliancy of conception and boldness of execution, perhaps he had no equal on either side during the civil war. Like other great and successful men he encountered the envy and jealousy of those less gifted and magnanimous than himself. "He was intensely patriotic and always willing to endure hardship and privation. His patriotism was of that intense kind that he would at any time have willingly sacrificed his life for the cause he served so brilliantly and well. His great courage, generosity, frankness, and patriotism endeared him to all the officers and men who served under him, and in every State of the Union they are now mourning his loss. "I spent some time with him at his home in New York three weeks ago last Sunday. He was then well, cheerful, and bright. He indulged much during the afternoon in reminiscence, and related a number of incidents of the war which I had forgotten. He mentioned a large number of mutual army friends who had died, and remarked: "'Gresham, we will join them soon.'" Ex-President Hayes paid this tribute to his military genius: "The only comparison of value that I choose to offer comes from abroad. We hear in regard to Sherman, from the French generals nothing but praise; from the German generals the same; from the English, General Wolseley speaks of him in terms that are altogether complimentary. Says Wolseley, however, 'Lee was a great general, and next to him was Sherman.' I would change the order. I admit for Lee a great character, accomplishments as a soldier and as a man, praise in every way except his unfortunate lack of wisdom. I do not now speak of motives, but of the military genius who was the military genius of the war. Place Lee where Sherman was. Place Sherman where Lee was. Place Lee at Chattanooga, even with Sherman's army. Would he have found his way to Atlanta, and at Atlanta cut loose from his base of supplies and entered upon the wild march for the sea three hundred miles away? I believe no man lacking the genius of Sherman would have entered on that march to
- 84. the sea. But come nearer home. Lee had the same opportunity, only it was ten times better than that Sherman had at Atlanta. Suppose Sherman had been in command of the army of Lee. Washington at that time lay completely in the power of an enterprising and daring commander, and with Washington captured, intervention from abroad would have come. I do not predict final defeat, for throughout all the action the finger of God was present, guiding and directing. I cannot believe that under any circumstances the cause of liberty and union could have failed, but at Washington was the chance of victory, and Lee failed to take it. More than that, he went to the Potomac, crossed it, and our disorganized army, without a commander, being divided between Pope and McClellan, was ten days behind him, and he marched on into Pennsylvania; and what did he do, and what would Sherman have done? Lee did not dare to lose communication with his base of supplies, and was driven back from Antietam with a divided army. Had Sherman been at the head of that army, and that distance between him and the pursuing forces, he would have gone to Philadelphia, Pittsburg, Buffalo, Cleveland, Cincinnati, and then cut his road back into Virginia. A little band of 4000 men under Morgan went through Ohio and Indiana, and Lee, with his great army, with nothing before him but wealth and supplies and cities able to pay tribute for not being burned, is not to be compared with Sherman." General Slocum said: "I have been acquainted with General Sherman since the beginning of the war. I first met him at Bull Run and afterward in the West, when my corps was sent there to reinforce Rosecrans. At that time he was tall and angular and his general appearance was much the same as it was in later life. My services with him began just before the capture of Atlanta. In that campaign the minutest details were attended to by General Sherman himself. Details as to the exact amount of ammunition to be taken by each corps, the exact amount of stores of each and every kind, were specified in his orders. During the campaign he alternated between General Howard and myself, riding with General Howard one day and with me the next. He was a great and most interesting
- 85. 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