This book "supplies a vocabulary of English words and idiomatic phrases 'arranged … according to the ideas which they express'. The thesaurus, continually expanded and updated, has always remained in print, but this reissued first edition shows the impressive breadth of Roget's own knowledge and interests."
Emmorey, Karen (2002) Language, Cognition, and the Brain: Insights from Sign Language Research. Lawrence Erlbaum Associates, xviii+383pp, paperback ISBN 0-8058-3399-4 $39.95
Reviewed by Zouhair Maalej, Department of English, University of Manouba-Tunis, Tunisia
Language, Cognition, and the Brain (LCB) includes an Introduction, eight chapters, an Epilogue, two appendices (one on handshapes in American Sign Language (ASL), another on communication forms in Nicaragua), an impressive bibliography (40 pages), and two indices (author and subject). True to its subtitle, LCB is about what sign languages can teach us about the nature and properties of language, the structure of cognition, and the workings of the brain. LCB is so rich with reports of experiments that the current review will appear almost reductionist.
BOOK'S PURPOSE AND CONTENTS The main thrust of the book is given in the Preface, which is the investigation of the key role of sign language in the study of language, cognition, and the brain in general. A synopsis of the chapters contents is also briefly spelled out.
1. Introduction Sign language is stated to be the backdoor to language, cognition, and the brain. Emmorey addresses some of the fallacies about sign language, namely that, (i) there is a universal sign language, (ii) sign languages are made up of gestures and mimes, (iii) sign languages are based on speech, (iv) sign languages can hardly serve the same subtleties and semantic complexity speech serves. Emmorey also brings evidence from Nicaraguan Sign Language (NSL) to bear on whether sign language develops as a smooth process or a discontinued, abrupt one, and whether it is the child or the adult who impacts the process, concluding that the birth of NSL is a discontinued, abrupt process, where the child plays a leading role. Emmorey wraps up the Introduction by pointing to the sociolinguistic dimensions of ASL, namely, geographic, gender, register variables, and the differences between ASL and other non-natural languages.
2. The Structure of ASL: Linguistic Universals and Modality Effects The Structure of ASL is reviewed at the levels of morphology, phonology, syntax, and discourse. At the level of morphology, ASL is said to differ from spoken language (SpL) by being noncancatenative. However, like for SpL, ASL word formation is guided by aspectual inflections, compounding, and derivation. The ASL lexicon is characterised more by iconicity than by arbitrariness, owing to the richness of the visual-gestural modality as against its auditory-vocal counterpart. The lexicon also includes English and foreign language borrowings and classifier predicates, which are fingerspelled and in violation of formational constraints of the lexicon, respectively.
Although it seems paradoxical to talk of a phonological level for ''soundless'' languages, ASL is argued to comprise three basic phonological categories: handshape or hand configuration (i.e. whether the palm is open, fist closed, or fist closed with the index finger pointing, etc.), location or place of articulation (i.e. whether the sign is made at the upper brow, the cheek, the upper arm, and horizontally or vertically, etc.), and movement (i.e. whether the path is straight or arced, and whether there is local movement, etc.). To ascertain a linear segmental structure for signs, location, movement, and handshape are likened to consonant, vowel, and tone, respectively. Handshape change is found to be coordinated with sign syllable rather than with the sign as a whole. Through a discussion of compounding, it was shown that the feature geometry rule found in SpL also applies to ASL. Spoken language and signed language are contrasted with regard to speech articulators, with SpL having one (i.e. the tongue) and signed language two articulators (i.e. the two hands, which either satisfy the symmetry condition or the dominance condition). Non-manual expression (such as mouthings and mouth gestures) is said to correspond to prosodic features in SpL.
At the syntactic level, ASL shows a SVO word order, whereby words may be topicalised and subjected to the same syntactic constraints as SpL. ASL also shows the grammatical dimension of nonmanual signs such as facial behaviour ((eye)brow patterns, lip patterns, mouth gestures, etc.). Pronominal signs are said to be compositional, and work by associating referents with a location in signing space when they are physically absent. But such pointing in signing space may give rise to ambiguity as to whether reference is to the entity in space or the location of the entity, which ambiguity is said to be resolved by reference to environing discourse.
At the level of discourse, turn-taking in ASL is regulated through eye gaze, but gestures in signing space are also used as attention-getting devices for a turn (hand raising, head nodding, etc.). Refusal to relinquish the floor is signalled through an upturned palm. Since the modality is visual, overlap between two participants may occur. Narrating in ASL marks reporting through ''referential shifts'' or ''role shifts'' (breaks in eye gaze with the addressee, possible shifts in head and body position, and changes in facial expression). Emmorey (1999b) argues that signers also produce ''component gestures'' (i.e. communicative gestures that are embedded within the utterance) and ''deictic gestures,'' which alternate with signs, thus indicating that gesture and sign are not co-expressive.
3. The Confluence of Language and Space This chapter is focused on how classifier constructions (CCs) are used in ASL to represent physical space and conceptual structure. CC express motion (The car meandered up a hill), position (The bicycle is next to the tree), stative-descriptive information (It's long and thin), and handling information (I picked up a spherical object). CCs in ASL are said to include (i) whole entity classifiers, (ii) handling and entity classifiers, (iii) limb classifiers, and (iv) extension classifier for handshapes. Whole entity classifiers can combine with movements such as position morphemes, motion morphemes, manner morphemes, and extension morphemes. Handling and entity classifiers can combine with some position, motion, and manner morphemes. Limb classifiers can only combine with manner morphemes. Extension classifier handshapes can only combine with extension movement morphemes.
As CCs received ample morphological and semantic treatment, Emmorey proposes to study them from syntactic and phonological perspectives. To address the question of whether CCs in ASL have gradient or categorical properties, Emmorey found that, unlike SpL which expresses spatial information categorically via prepositions and locatives, ASL expresses spatial information via categorical, gradient, and analogue representations. Another possibility for ASL is to combine two classifier predicates to express simultaneous constructions. ASL signers are said to adopt intrinsic, relative, and absolute frames of reference within signing space, with the possibility for intrinsic and relative frames to be expressed simultaneously. Emmorey (1999a) reports on evidence whereby the right-hemisphere is involved in processing spatial information.
Time lines are based on the FUTURE IS AHEAD, PAST IS BEHIND, and PRESENT IS AT THE BODY conceptual metaphors. Deictic time lines (next, tomorrow, yesterday) are signed with reference to movement away from the signer's body, drawing mappings between signing space and temporal structure. Such times refer to points in time related to current discourse. Anaphoric time lines proceed diagonally, and are used in comparisons or contrasts of time periods within discourse. Sequence time lines proceed from left to right, and refer to the order of events in time. To encode other abstract concepts, signers are said to have recourse to two forms of mapping: abstract-to-concrete and iconic-to-linguistic. Such mappings are corroborated by Wilcox (forth. 2002).
4. Psycholinguistic Studies of Sign Perception, Online Processing, and Production One aim of this chapter is to investigate the amount of universalism and specificity in language processing and production. Signers are said to show categorical perception for hand configuration, which suggests that language experience affects perceptual categorization. Lexical access has been shown to be faster in ASL than for spoken English, owing to early recognition of simultaneous (versus serial representation for spoken English) hand configuration, place of articulation, and movement, and the scarcity of signs sharing initial hand configuration and place of articulation. In terms of lexical organisation, both ASL and SpL have been shown to follow a spreading activation model of recognition, whereby the decision time for the recognition of a lexical item is lower if semantic associates are presented (e. PENCIL as presented after PAPER) and vice versa. Online processing of ASL has been found to be similar to that in SpL, but with sign languages showing how spatial location can affect processing and representation in memory.
Owing to its visual-manual modality, sign production can offer two simultaneous constructions distinctly and dominance reversal constructions by right-handed signers with their left hands and vice versa (for backgrounding parenthetical information or foregrounding contrasts and comparisons). Sign production in ASL is not immune from prearticulatory and postarticulatory editing (a slight hand wave, palm outward, headshake), whose indices are self-interruptions and self-repairs. Pauses as indication of filled pauses like ''ums'' and ''uhs'' have also been noted among signers. Because signers do not see and hear themselves while signing, such pauses and editings may be substantially different from that for speakers who can hear themselves. Whispering (when others are discarded from one's talk by displacing it aside) and shouting (in crowded places or when the hearer is in a distal location) are common in sign languages.
Lexical retrieval suggests that semantic information is independent of phonological information. Slips of the hand in ASL may come from errors in hand configuration, place of articulation, or movement. The fact that stranding errors are rare in ASL is accounted for by the nonconcatenative nature of its morphological processes. The early repair of errors may also be a result of the same processes and the comparative slowness of sign articulation. The production of gestures among signers is also rare owing to the fact that both hands are usually busy; instead, contemporaneous body and facial gestures with signing are more common, and seem to have a communicative (rather than metaphoric and facilitative) function.
5. Sign Language Acquisition: Early development Evidence suggests that speech and sign acquisitions know the same regularity and developmental stages, suggesting that neural mechanisms are not modality-specific. Manual babbling (appearing between 10 and 14 months) is syllabic, and tends to be reduplicative. Signs are produced earlier than words owing to the development of the motor system and visual cortex as against the vocal tract and the auditory cortex. What seems to cause problems for baby talk is getting the right hand configuration. Motherese appears to exist in sign language, and to perform the same function as for SpL. However, because the same modality is involved when looking at adults signing to babies and attending to the spatial context in which objects are presented, it might be thought that language development is impeded by its very visual modality. Mothers are said to attune their talk to the child's communicative development to avoid these setbacks. The acquisition of syntax and morphology is almost the same in speaking and signing children. With very few exceptions, overwhelming evidence suggests that iconicity does not facilitate acquisition among young signers.
5 (cont'd). Sign Language Acquisition: Later development Fingerspelling is said to be understood and acquired by signing children even when it is rare in the linguistic input. Not until the age of five (or even six) do young signers manage to conquer signing space, which is an error-ridden and arduous task for them. Such late and arduous acquisition is due to problems relating to spatial memory, i.e. the difficulty for the young signer to remember ''multiple referent-location associations'' (p. 192). Further evidence that iconicity does not facilitate acquisition among young signers even in later cognitive development comes from the late acquisition (between 8 and 9) of classifier constructions owing to the complications in terms of handshape, movement, and location they pose to the young signer. Narrative development as signalled by referential shifts is a crowning development. Young signers are said to fail to signal through body gesturing and eye gaze the perspective from which the narrative is told, which suggests the hand-before-face development. And they seem to master referential shifts for quoting before constructed narrative.
6. The Critical Period Hypothesis and the Effects of Late Language Acquisition Early and late signers are said to differ in their spending cognitive effort solving (a) semantic and (b) phonological details, respectively. Late signers suffer also from morphological and discursive impairments, although some of their processing mechanisms for word order and coreference remain intact. Comparing ASL late acquisition and second language acquisition, Emmorey argues that ''the long-term effects of a delay in first language acquisition appear to be much more detrimental than the effects of acquiring a second language late in childhood'' (p. 217), the reason being that the learner of a second language has already the linguistic knowledge acquired with first language acquisition. However, deaf children not exposed to sign language develop home sign gesture systems (mostly iconic and ergative in nature), which have been observed to show regularity across children from different cultures.
Delayed acquisition is said to be followed or accompanied by cognitive delays rather than deficits. In general, deaf children with deaf parents have been observed to outperform deaf children with hearing parents. In particular, a lack of early language experience is said to affect short-term memory capacity (recall), which affection fades by adulthood. Deaf children with hearing parents are said to have problems with the development of the theory of mind, i.e. the capacity to attribute mental states (such as beliefs, intentions, desires, emotions, etc.) to people. Causes for such a delay relate to the absence from the child's input of conversation and syntax, which are thought to favour the development of mental states.
7. Memory for Sign Language: Implications for the Structure of Working Memory Evidence from SpL and ASL suggest that recall in working memory has a phonological rather than semantic bias. Recall of words with different meanings seems to be enhanced while recall of words with similar meanings seems to be impeded. Working memory span in signers is said to be shorter than that in speakers owing to the fact that signs take longer to articulate than words. Spatial coding can be used as a memory device: signs with a spatial neural basis are easier to memorise than bodily-based signs. Signers perform equally well in remembering signed numbers forward and backward, suggesting that signers' working memory is not unidirectional as that of non-signers. As summed up by Emmorey (p. 240), ''speech-based working memory appears to excel at using time to code serial order, whereas sign-based working memory is able to use space to code serial order.''
8. The Impact of Sign Language Use on Visuospatial Cognition Regarding motion processing, signers are said to be more expert than non-signers at detecting motion, which suggests that this capacity is enhanced among signers as a compensation for auditory deprivation. Neuroimaging shows that motion perception is left hemisphere-based. With regard to face processing, signers are said to outperform non-signers in face perception, and ASL signers are said to have enhanced ability to identify emotional facial expressions. Concerning mental imagery, signers are said to have enhanced performance on mental rotation tasks owing to their experience with ASL. Signers have also been found to have a broader spatial memory span as compared to non-signers. Emmorey (p. 267) argues that there are capacities (motion discrimination thresholds, face recognition ability, maintenance of visual images, localization skill, visuospatial constructive abilities, etc.) in which signers and non-signers do not show differences, which suggests that these capacities do not depend on sign language use but relate to general cognition.
9. Sign Language and the Brain Signers with left-hemisphere damage were observed to perform worse than signers with right-hemisphere damage, which ascertains the location of sign language in the left hemisphere and the importance of it in signs comprehension and production. With the exception of the puzzling study by Neville et al (1998), many experiments having to do with brain activity corroborate this finding: cortical stimulation of the left hemisphere in the surgical treatment for seizures, the Wada Test injection, visual hemifield presentation, dual task paradigm (for both signers and non-signers), event-related brain potentials, positron emission tomography, functional magnetic resonance imaging, etc.
To further ascertain the specialisation of the left hemisphere in language comprehension and production, Emmorey mentions many other experiments: dissociating sign language from symbolic gesture (some sign impairments versus well-preserved gesture or pantomime), dissociating neural control for motor versus linguistic processes (capacity to copy meaningless movement sequences versus linguistic impairment), dissociating sign language ability from non-linguistic spatial cognitive ability (right hemisphere damage resulting in non-linguistic spatial cognitive deficit of, e.g., the visuospatial abilities, but not in sign language aphasia), which should not entail that the left hemisphere, as Emmorey rightly points out, is aspatial.
The neuroanatomy of sign language production is thought to be the same as that for SpL: the Broca area. The neuroanatomy of sign language comprehension depends on the temporal lobe independently of the modality, with the superior temporal gyrus undergoing functional reorganization in signers to receive visual sign input instead of auditory sign input. The neuroanatomy for the production and perception of emotional and linguistic facial expressions seems to involve the right left hemisphere. In the same way the left hemisphere is not aspatial, the right hemisphere is not alinguistic. The right hemisphere is said to play lexical (when words are imageable or have concrete referents) and discursive roles (when coreference is misused). Damage to the right hemisphere is said to cause disruption to the organisation of signing space.
CRITICAL REVIEW The book is a comprehensive document about how ASL is used by signers to function in space and mean. Findings in ASL have been evidenced with experiments mainly done by the author herself, with associates, or by other researchers. The book is also sprinkled with cross-linguistic references to, and evidence from, British, Chinese, German, Japanese, Jordanian, Swedish, etc. sign languages to corroborate ASL. Each chapter of the book includes a short concluding section (and sometimes implications), wrapping up the main findings. The book is well indicated for students of SpL and sign language alike, and a landmark for researchers in sign language at large and especially its neuropsychology.
But beyond its descriptive (mainly) and argumentative dimensions, the book contributes insights for cognition and language at large:
(i) Linguistic prewiring principle Evidence from sign languages show that constraints on syntactic form (e.g. structure dependency) are universal across modalities (hearing and visual), which suggests that ''we are born with the ability to acquire language'' (p. 204). This is a confirmation for what is known as the innateness hypothesis.
(ii) The poverty-of-the-stimulus principle Evidence from sign and SpL, whereby the role of the input in language acquisition is restricted to a minimum intervention from adults, constitute confirmation of the barley-distillery-whiskey metaphor for Chomsky's ''poverty of the stimulus hypothesis.'' As Emmorey pointed out (p. 8), ''over 90% of Deaf children are born to parents who do not know sign language,'' and yet all of them come with some form of sign language to communicate with.
(iii) The child contribution principle and the critical period hypothesis Evidence from sign language corroborate the fact that late language acquisition not only impacts the level of mastery of grammar but also affects language comprehension. Early native signers have been observed to outperform their late native counterparts. However, home-signing among children reared in an environment without signs input suggests the intervention of the child to develop a transitory system of communication even in the presence of a very impoverished adult input. The fact that both speaking and signing children subscribe to a critical period of language acquisition is evidence that the process is neural, and is, therefore, modality-independent.
(iv) Design features of language: arbitrariness and/or iconicity Evidence from sign languages suggest that, although iconicity does not appear to play a facilitative role in early acquisition and in adulthood, arbitrariness of word forms does not seem to be ''a basic or necessary characteristic of human language'' either, and that it is ''the articulatory and perceptual resources of spoken languages [that] limit the iconicity of spoken words'' (p. 17).
(v) The uniformity of language acquisition principle Evidence from both spoken and signed languages point to the uniformity requirement (i.e. the fact that all children within the same culture and across different cultures go through more or less the same processes acquire a language). However, differences in uniformity between spoken and signed languages are due to ''linguistic and cognitive principles that are not specific to the visual-manual or aural-oral modalities'' (p. 204).
(vi) Right versus left hemisphere specialisation Evidence from sign language shows the involvement of the left hemisphere in motion perception. A parallel study of gestures and hemispheres by Feyereisen (1999: 8) indicates that meaningful gestures activate the left hemisphere whereas meaningless ones activate regions in the right hemisphere, although ''the systems for gesture imitation and recognition are not independent.'' Motion that is linguistic is meaningful, therefore activating the left hemisphere. Motion that is linguistically irrelevant for signing is relegated to the right hemisphere. This division of labour between hemispheres, whereby the left hemisphere specialises in linguistic matters and the right hemisphere in non-linguistic matters, seems to be confirmed by sign language research, although studies of neural disorders by Corina (1999: 29) suggest that deficits are the result of ''lesions more posterior than those observed in users of spoken language'' in the left hemisphere.
A point needs to be made concerning the attribution to ASL of merits such as the capacity to identify facial expressions, performing well in mental rotation, the ability to generate mental images, etc. Although Emmorey made it clear in many places what the advantages of using a sign language are, perhaps, as Emmorey pointed out in the Epilogue, comparative studies with ASL could be used to tell us whether such performances of ASL signers are owed to using a sign language at large or depend on using ASL in particular.
Finally, a very small set of typos should be signalled: to chose (instead of: to choose)(p. 83), more that (instead of: more than)(p. 174), do not to expose (instead of: do not want to expose)(p. 206), may decline due the process (instead of: may decline due to the process)(p. 218).
REFERENCES Corina, David P. (1999). ''Neural Disorders of Language and Movements: Evidence from American Sign Language.'' In: Lynn Messing & Ruth Campbell (eds.), Gesture, Speech, and Sign. Oxford/New York: Oxford University Press, 27-43.
Emmorey, Karen (1999a). ''The Confluence of Space and Language in Signed Languages.'' In: Paul Bloom, Mary A. Peterson, Lynn Nadel & Merrill F. Garrett (eds.), _Language and Space_. Cambridge/London: The MIT Press, 171-209.
Emmorey, Karen (1999b). ''Do Signers Gesture?'' In: Lynn S. Messing & Ruth Campbell (eds.), Gesture, Speech, and Sign. Oxford/New York: Oxford University Press, 133-159.
Feyereisen, Pierre (1999). ''Neuropsychology of Communicative Movements.'' In: Lynn S. Messing & Ruth Campbell (eds.), Gesture, Speech, and Sign. Oxford/New York: Oxford University Press, 3-25.
Neville, H., D. Bavelier, D. Corina, J. Rauschecker, A. Karni, A. Lalwani, A. Braun, V. Clark, P. Jezzard & R. Turner (1998). ''Cerebral Organization for Language in Deaf and Hearing Subjects: Biological Constraints and Effects of Experience.'' Proceedings of the National Academy of Science 95, 922-929.
Wilcox, Sherman (forth. 2002). ''The Iconic Mapping of Space and Time in Sign Languages.'' In: Liliana Albertazzi (ed.), Unfolding Perceptual Continua. Amsterdam: John Benjamins Publishing Company.
ABOUT THE REVIEWER:
The reviewer is an assistant professor of linguistics. His main interests include cognitive linguistics, metaphor, pragmatics, cognition-culture interface, pragmatics-cognition interface, neuropsychology, psycholinguistics, critical discourse analysis, stylistics, sign language and gesture, etc.