The English Tone of Voice: Essays in Intonation, Prosody and Paralanguage by David Crystal Review by: Harvey Sarles

Language in Society, Vol. 8, No. 3 (Dec., 1979), pp. 425-439 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/4167094 . Accessed: 07/03/2012 09:44
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Language in Society.

http://www.jstor.org

Lang. Soc. 8, 425-475. Printed in Great Britain

REVIEWS
MODES AND CODES The English tone of voice: Essays in intonation,prosody and DAVID CRYSTAL. paralanguage. London: Edward Arnold, 1975. PP. I98.
David Crystal's book on 'Intonation, prosody and paralanguage' is at once its own paradox, a critical labor of love, and a thorough, densely-argued set of claims about the state of the art. It is, in some senses, so many things that it is almost impenetrable unless the reader comes to it with a well-developed set of puzzles about the nature of language. It is already 'beyond' linguistics because Crystal wants to account for more of speech or of the stream of speech than is usually considered interesting or necessary. And it is an attempt to be thorough from at least the point of view of a linguist. The book is a set of essays attempting to 'take a stage further the approach to the analysis of non-segmental phonology first described in Prosodic systems and intonation in English' (vii), Crystal's earlier (I969) book. He felt this earlier treatment to be inadequate on a variety of grounds, theoretical and practical. This volume is thus calculated to deepen and broaden approaches to those phenomena which occur vocally, including most of those which are not ordinarily considered to be language. As the introduction points out, 'the phrase "tone of voice" makes a succinct and perspicuous title, but by itself is inadequate' (vii). Terms such as 'nonsegmental' or 'suprasegmental phonology' relate more to technical linguistics, but are 'negative'. The trouble is that linguistics has decided (following an essentialist dualism) that only some aspects or features of speech are linguistic. This book talks about what else there is and what is 'left over'. Crystal separates the field pragmatically - much as different interests approach what may (or may not) be identical phenomena: intonation (Chapters I, 4, 7); paralanguage (2, 3); functions of 'the whole range of prosodic and paralinguistic features' (5, 6, 8). He attempts to review the literature and bring some critical thought to bear on . . whatever. This is a brave attempt, and more complete than I would have thought possible. Once one accepts that the total range of speech must be somehow accounted for, one is in a useful position to criticize linguistic theory. Chapter i, 'Prosodic features and linguistic theory', is a forceful criticism of current linguistics - I think, of any school. Chomsky is the principal whipping boy, probably because of his quality and claims. But one suspects that he is a surrogate for almost everyone in this century and before. Crystal's major criticism is that by distinguishing what is claimed to be 'linguistic' (always on dubious grounds), one 425

LANGUAGE

IN

SOCIETY

tends to oversimplify whatever else there is in verbal behavior, into a singular phenomenon, usually 'intonation',and to use a single variable- usually sentence 'stress', or tonal center - also called 'tonicity' - to account for all the residual problems of linguistics. (The naturalist is reminded of the same move in separatingthe human species from all others.) Crystalpoints out that intonation'is not a unitary,homogeneousphenomenon' (ii) - even if everyone talks about it as if it were. This oversimplification results especially from Chomsky'spartialexaminationof 'only a small set of sentences' (ii), trying to account for some sorts of interface or relationshipbetween deep and surfacestructures,presumablyof a type which interpretsemanticstructuresan importantidea for Crystal. Here the phonological,basic componentsare split into segmentaland non-segmental components. Instead Crystal claims that intonation is the 'product of the interaction of features from different prosodic systems' (ii). He posits a 'partly hierarchic organization, such that the basic unit of prosodic organization,the tone-unit' (author's italics), has certain featureswhich he laid out in full in an earlierwork and elaboratesin this book as well. He proceeds to examine the 'main structural characteristics'of this model in Chapter i. As he points out, this approachhas the merit of showing that intonation is much more complicated than has been thought by Chomskyet al., irrespectiveof Crystal'sworkingout of details. Over and over, the author points out that one must come to this study using a corpus of spontaneousconversation- but, of course, the reasonfor underestimating the complexities of intonation have to do with deep assumptions about the 'location' of language phenomenaand problems. Crystal then spends some 30 pages, still in the first chapter, laying out an elaboratesystem which is built on taking-as-givenmeaning phonetic and substance: 'the question is to determine what components are necessary in order to interrelatethem' (i 2). While one appreciatesthis attempt,it smacksof certainfamiliar dualisms (e.g. mind-body), in which each component of the equation remains unexamined while we are told to direct our attention to the interface,for the putative solution. The system is a typology built on some notion of 'tone-unit' (I5) which occupies/is assigned to a sentence or phrase.There is a notion of basic pattern, 'subject+verb+complement and/or object, with one optional adverb, in this order'(i6), which has a single tone-unit. Syntacticstructureand variouslevels of prosody work somewhat independently,but hierarchically.In Crystal'sterminology, a clause is 'subjectedto a prosodicoperationwhich gives it an intonational identify' (17). This is then fed into the next level, etc. Obviously, one has to assume something basic about what levels are, how they are ordered, what expansions follow from which, what can be 'accounted for' and what cannot, and what it means to predict. Using his system to analyzeconversation,Crystal claims: 'Out of the I2,000 tone-units examined, about ioo were incapable of 426

REVIEWS

prediction from the above rules' (2I). This reviewer feels incapable of judging the meaningfulnessof this claim in the form in which it is stated. It seems not difficult to make 'predictive' statements in this fashion. The question is: does this sort of prediction have much to do with what will go on in some other conversation?Do we know how to attend to the data in interesting, possibly correct ways? Crystal goes on to point out that it is necessary to have already done a great deal of prosodic categorization(22) before one can specify, or even raise, questions of tonic placement. He writes out how to do this, briefly, by a sort of 'contrastivesemantics'.Like most definitionaltypologies, it seems to presumea great deal of what ought to be discovered and smacks of being deeply tautological. Crystal's final theoreticaldiscussion in this ambitious first chapter, is to discuss the nature of 'tone'. He wants to consider tone 'in terms of the interaction between two systematicallydistinct features, syllablepitch range and tone' (28). The authoris awareof (some of) the difficultiesin this arena,attributingthem to rough, but underlying, categories of: 'context, system and connectedspeech' (29). He deprecatescontextand likes system. In his discussion of context, one finds the clear appearanceof the Platonist, because: 'Statements referring to the importance of context in intonation analysis . . . reflect a movement away from the view that a nuclear tone . . expounds a single, "basic"meaning,alwayspresent regardlessof context (though contextual variations may add certain overtones), to a view that there is no common meaning underlying all instances of its use. . .' (29). One can almost taste Crystal's hope for the salvation of basic, constant meaning. If it is not in prosody, where then is it? It is almost as if the quest for the nature of knowledge must stop here - the bottom line. In his defense - for Crystal is, finally, knowledgeable and reasonable- he does acknowledge that this is a quandary. His fear is that 'context' will become the latest fad in residual explanatorydevices, merely placing the problems out of our line of vision yet again. Instead Crystal outlines a more 'manageable' set of five variables which, briefly, call for close attention to approximately all the behavior- including facial expression, kinesics, etc. - in which a phrase is embedded. Why this is something differentfrom context, or whether it is context-reduced to 'manageable proportions',is not easy to derive from the author'sdescription.Those who delve seriously into languageand facial expression have not yet found much to be manageable. This sort of attempt to 'reduce' complications to an arena in which one has little experience,as if it is ultimatelysimple, is a common ploy but not very helpful. Less explicitly a theme, but not inobvious, are references to affective/emotionalstates which the author takes to have some kind of meaning, whose examinationwill clarifythe problemsof intonation.One is certainthat this is correct in some sense, but Crystal's reasoning seems not to be directly on target. This is the sort of problem which should lead scholars to cast their nets 427

LANGUAGE

IN

SOCIETY

more widely into arguments about human nature. Crystal's review of this crucial area appears at once overly optimistic and simplistic, and will likely lead to new impasses once the complexity of human bodies in interaction is realized. The crucial problem with this sort of scheme is that Crystal and, one guesses, most linguists, believe that they are good and consistent observers of 'prosodic features'. Linguists, being insiders, can be quite consistent observers, but one has severe doubts that we are very good - being part of our own ongoingness. The problems of observation seem to be difficult and a leading reason why questions of human nature are now being seriously considered. Chapter I - a full 46 pages - thus completed, much of the remainder of the book might be described as 'critical reportage' with occasional theoretical excursions into particular puzzles (especially Chapter 4). Chapter 2 is entitled 'Current trends in paralinguistics'. Beginning with the I962 Indiana Conference on Paralinguistics and Kinesics (later published), the 6os developed in three ways: (i) Tragerian (I958); (2) non-Tragerian; and (3) confusionism. Why did paralinguistics drift into obscurity? Crystal states: 'When a theory (in this case, generative theory) is so much in the ascendant, any earlier linguistic approach which does not take cognizance of that theory, and attempts some comparison with its own claims, is necessarily going to distance itself from the eye of the majority of linguists and become, in effect, of historical interest' (48). Trager and those of his students who remained 'loyal' to his interests, instead of discussing paralanguage within the confines of generative theory, became obscured. Few of us even remained with Trager's puzzles - intriguing as they were and most people 'converted' to 'modern' linguistics. The pressures were severe. The author claims that none of the Tragerians have answered the criticisms levelled at the approach of Trager and Smith. One of the major reasons, of course, is that during periods of theory ascendence, one does not get taken very seriously unless one jumps on the bandwagon. One who stayed with Trager's puzzles would have found few intellectual grounds for mutual conversations with linguists in the ascendant phase. Other reasons for lack of interest in paralinguistics are reflected in lack of mention in standard textbooks during this era. The linkage with semiotics has begun to repopularize it, but likely from quite different and non-linguistic perspectives. Crystal claims that there was noformal theoretical basis in Trager's essentially descriptive formulation of paralanguage. I concur, but hasten to point out that there are deep, probably unreconcilable differences between linguistic formalists and field-oriented descriptivists concerning rather basic issues: what a theory is or should be, 'where' the data is 'located' (in the observer's experience, in the informant's mind, in his conversations, in spontaneous speech); what 'prediction' means; what is an 'accounting' for; etc. As he goes on to say, it remains unclear what paralanguage is: '. . . the quite remarkable range of subject-matter which is allowed under the heading of 428

REVIEWS

paralanguage (5x).' There is somewhere a champion for every imaginableview, either in the literature,or elicited in Crystal's correspondences.He finds seven distinct 'basic' senses, which most scholars use, and suggests that there is still some (vast?) distance between these and science.
(i) Human and non-human. Why anyone would seriously consider (human) linguists discussing animal vocalizationis surprising. The similaritiesare trivial to Crystal. (2) Including non-vocal as well as real features of human communication kinesics. He finds this unlikely: verbal is verbal. (3) Including all non-segmental ('suprasegmental')features and some segmental ones. He sees this as a residual category, a dumping ground for everything which is consideredto be non-linguistic. Later in this volume he considers this at length. (4) Including voice quality; i.e. person-identifyingvocal characteristics.He opposes this on the grounds that some voice qualities are 'biological',which for Crystal means uncontrolled, non-contrastive behavior, aspects of one's being (and somebody else's subject matter one imagines). (5) Including only non-segmentalfeatures, but excluding prosodic phonemes and voice quality: the dominant sense of the term, for Crystal. Here 'pitch, stress and juncture phonemes are considered linguistic, and any other suprasegmental effects are considered paralinguistic'(54). (6) Including only a sub-set of non-segmental features other than prosodic phonemes and voice quality. This has been restricted to 'non-segmental phonology for a particularlanguage' and needs to be more generalized, thus comparative.(Crystal's criticism of some of his own earlierideas.) (7) Functional definitions. This would include, possibly, all those aspects of vocalizationwhich have to do with affect,social-groupidentification,personality, etc. Most of them are probablynot structural,in a linguistic sense, but .. .?

This list illustrates the vast range of notions which someone considers to be paralanguage,and about which there is very little agreement among scholars. There has been very little experimental work, few descriptive studies - and most of this in English. By the end of the second chapter, the reader begins to sense the need for some guidance into this wonderland. Part of the problem seems to stem from underlying theories about language, at least about what is 'linguistic'. Some scholars had hoped that meaning is somehow specified or located in paralanguage. One senses that they do not really want to know how or where, however: 'context decides'. Crystal disagrees with Trager on the basic issues of whether paralanguage is linguistic and what this decision means. For Crystal, these are argumentsabout language as a whole, and what the ' "proper" subject-matter of linguistics'
429

LANGUAGE

IN

SOCIETY

is (59). The 'segmentalprinciple'is 'a majorprejudicein linguistics'.He believes that all vocal effects have to be considered 'in the first instance'. His solution to avoid the arbitrarinessof the splits-to-come between what is systematic and/or discrete and what is not is to introducethe concept of a 'SCALE of linguisticness'. Only the vocal effects 'lackingany semanticforce would then be considered non-linguistic' (60). Crystal admittedly uses this notion more for the 'attention it focuses on the need for analytic criteria to be made explicit' (6I) than as a fully serious alternativeanalysis.He is criticalof the largeamountof 'busy-work' which has been spent on proceduralmatters, where no serious thought has been roles of given to what the analysis is about. Functional and socio-paralinguistic phenomenaare at once obvious and inadequatelyspecifiedso far. Clearlyone has to be carefully critical in reviewingstudies which purportto be in this area. is The third chapteraddressesthe generalbelief that paralanguage some direct reflection of the emotions. Its title is: 'Paralanguagein animal and human communication'.Since - it is often said - humans and animals share the same/ similar emotions, tone-of-voice is an areaof overlapbetween the species. Crystal thinks that this is either wrong or premature. He bases his argumenton his 'broad' definition of language- including nonsegmental phenomena as long as they are 'meaningfully contrastive' (65). He will exclude kinesics and other non-verbal phenomena 'until such time as it can be shown that there are sufficient parallels between paralinguisticand kinesic structuresto warranta conflation(65).' This is a deep and thorough statement about what is deemed to be human, assuming 'language'to be the differentialforce. I distrust it as being incapable of convincing Crystal or others that there are 'sufficient parallels', because the very distinctions between human and animal are essentially the same as between whateverwe choose to call paralinguistic(read 'linguistic' for Crystal) and kinesicstructures.This, as I have come to understandit, is merely a modernized form of the underlying dualism and can shift to include/excludewhatever we choose, and what is 'selling'. Example: pretty flimsy grounds. Crystal states (having clearly denied in the 'Now that a great previouschapterthat we yet know much about paralanguage): deal more paralinguisticstudy has taken place, it is possible to re-evaluateits status vis-a-vis human and animal communication,and to conclude that paralanguage is much closer to the rest of languagethan was originallyanticipated. The overlap with animal communicationis minimal and trivial (66).' All this says - on purely definitionalgrounds - is that animalsdo not have meaningand is yet another form of the sign-symbol dualism. Nothing new here. If there is anything to the possibility of human-animal comparison, Crystal appears uninterested in hearing it, by whateverform of demonstration. Chapter4 raises a single(?) issue, and attempts to show that the grounds for yet another general belief are sticky at best. The question concerns 'Relative 430

REVIEWS

and absolutein intonationanalysis'(74). Here is Crystal,thoughtfullywondering why everyone says that pitch is relative to other pitches in a pattern. What has one bought? What does this really involve? First, he points out that some properties of intonation systems could be relativistic, and others not. He states that the claim that 'intonation is relative makes sense only if some kind of absolutism is introduced from the very beginning' (75). I think he is right, and wish for more elaboration. Secondly, he states that the concept of absolute pitch should not necessarily be tied to its acoustic definition. Taking into account any person's 'voicetype', there can still be a great deal of consistency in anyone's speech, which within some notion of absolute. He consistency is in many senses understandable argues persuasively for this weaker notion of absolute. I would agree with this on the grounds that there is little actual confusion in what people hear in spontaneous conversation: thus there must be some sorts of grounding or consistencies. He then turns his attention to what the term 'relative' entails. There is, for example, no infinite range of pitches - or if there were, we couldn't hear them anyway. 'We must ask exactly what the principle of relativityis being invoked to explain' (77). This turns out to be a kind of criticism of tonal phonemes, and Crystal persuasivelyargues that the relativity hypothesis can work only in two ways: (i) that 'we clearly answer the question, "relative to what?"' and/or (2) 'that we can postulate an absolutely defined pitch level (or more than one), to which pitch variationscan be related independent of context' (79). In either case there must be some sort of factors or relationshipswhich have a kind of consistency, if not an obvious constancy. Crystalpoints out, for example, that vocal impressionistscan use some limited vocal correspondencesto real dialects, which nonetheless serve to 'convey' that dialect. It certainly works, and is likely similar to how we 'see' individual and group differences- i.e. by some sorts of stereotypingprocesses. We alreadyhave some perceptualgrid in mind, and actively match or interpretwhat we hear with this filter. The 'absolutism' Crystal postulates may/must be in the hearer's interpretiveframeworksas well as in the vocal performance. This causes Crystal to conclude that linguists have, in their collective minds, some sorts of absolutist notions which they do not make explicit. (I would go further and claim that many linguists actively deny having any such notion, making them less susceptible to persuasion by demonstration or argument.) They argue by casual 'reference to the "middle" of the voice-range' (8i), or similar working distinctions. There are no good reasons to argue against an absolute level of intonation, and several reasons why this is useful, and probablytrue. 'At the very least ... an unqualifiedrelativisticview of intonationis just as unfavorableas an unqualified absolutist one ... that a blend of both notions is required... I would 43'

LANGUAGE

IN

SOCIETY

propose a model in which pitch relativityis constrainedby the absolute levels, (83). . .. at least three pitch reference-areas' I agreewith Crystal,and have wonderedat length why this is inobviousto the about language. It has, in my view, a great deal to do with the observer/thinker sorts of 'ongoing' speakers we are, and how to 'break out' of our ongoing contexts. The next chapter (5) delves into the fact that different(sorts of) people speak Research differently;that some of these differencesare prosodicor paralinguistic. has, however,been 'almostexclusively based on the study of lexical, grammatical and segmental phonetic characteristics. .. illustrated solely with reference to restricted usage items of vocabularyand of grammaticalinflectionsor structures, and to differences in the articulationof vowels, consonants, and vowelconsonant sequences' (84). While there have been some recent moves in the directions of Crystal's interests, especially by Hymes, there has been little 'detailed account of the non-segmentalphonology involved: the referencesstay at a maximallygeneral level' (85). In this chapter, Crystalattempts to 'outline this area of study', and to survey some of the extant literature.He divides the main referencesinto 'five generally recognized categories. . . sex, age, status, occupation and functions (genres)' (85). Under sex are included references to informal/anecdotalreports about 'clucking like old women', having a 'sexy' voice - the lisping 'effeminate'voice, huskiness, registerswitch. This happens differentlyin different languages. Most of the age referencesin the literatureare to 'baby-talk',but its nature is left to everyone's(possibly different)experience.While it is apparentlyobvious to everyone that there are age differencesin speech, these remainundescribeddifferent scholars apparently differing on whether they are (even potentially) observable. This is also true of status- different sorts of people sound like one another; but differindividuallyand as groups from others: social classes, formal-informal, etc. Here it is important to study the 'ideals of speech behavior used . .. the systematic examinationof elocution handbooks... in order to understandhow differentpeople perceivedialects'.And of occupation commentslike 'you sound like a clergyman' represent some sorts of facts-of-life. While there were in psychology, from about 1935-50, attempts to demonstratethese patterns, they were not satisfactoryon a variety of grounds. This is not merely commentaryabout paralanguage,but a statement about we the sorts of contextual interactor-observers are. As ordinary speakers, we carrywithin us some sorts of comparativeframeworks.We are dialect comparators in our everydaybeing, an ambience not particularlygiven only to linguists. Linguistic perception varies individually, but also from the shared perceptual viewpoints of possibly large groups of persons who share one or more of the (dialect?)factorslaid out in this chapter.Not to takethese into accountin treatises 432

REVIEWS

on human language seems either careless or silly from the vantage of Crystal and other thinkers about language as occurring in the real-social world. There is, of course, a long untold story about the senses in which this is inobvious to the dualist whom Crystal is taking to task. Crystal cites reported instances of a surprisingly large variety of languages in which paralanguage and social factors come together in interesting ways. This largely impressionistic commentary 'has been marred by lack of an adequate theory, inexplicitness of definition and certain methodological weaknesses (92).' Nonetheless social anthropology may benefit to a large extent by ideas and methods derived from studies of paralanguage. Unaware of the potential subtleties of speech in every language, social anthropologists are often overly literal when they believe they are dealing with an 'exotic' language, in the reviewer's experience. In order to make us more aware observers, Crystal lays out a brief typology derived from his and Quirk's earlier (I964) work (i9). He claims that they account for all of the sorts 'of effects noted in this chapter, and groups them into systems on the basis of shared formal characteristics' (92). He distinguishes several (necessary) 'conceptual stages'. To this reviewer, the danger of such a typology, raised to the level of something conceptual, is that those who have not worked within Crystal's system may take this system to be the newest form of cookbook or laundry list, and faithfully record what they take to be instances of what they think they should hear. His 'conceptual stages', as they appear in lineal form, are as follows.

Non-linguistic vocal effects (A) VOICE QUALITY - the 'background' which the analyst uses to recognize the person or groups, as distinct. Here, I agree with Crystal's idea of voice quality, but think that it is more important for the analyst to find out how the hearers make this distinction, rather than - at least in addition to - how the analyst does. This may make a very large difference in 'where' we take the subject matter of linguistics to be 'located' and whose orientation is correct. Thinkers such as Crystal and this reviewer may find ourselves quite incompatible in the long run, and the deep difference is likely to be found in the assumed proper role of the analyst. (B) PHYSIOLOGICAL REFLEXES - 'Physiological reflexes, such as coughs, sneezes, or husky voice due to a sore throat, may also occur with speech, and must also be discounted as background "noise"' (93). I disagree strongly! What is a reflex? Does it have 'a life of its own'? It might easily be message, not noise. If something in an interaction 'alters' one's physiological being, then a cough or sneeze may be communicative. At the least, such an occurrence gives the interactor some (extra) information about the speaker's state. At least some physiological 'reflexes' 433

LANGUAGE

IN

SOCIETY

are 'catching' - e.g. yawning, dry throats. Inferring how Crystal gets to a statement such as this, it would appear that he does believe in a deep, independent aspect to our being which has its own existence. This seems highly debatable. Semiotic frame - this seems to be a residual category including 'everything else'. Why is this a 'conceptual stage'? Vocal-auditory component (A) SEGMENTAL-VERBAL - linguistics-so-far, plus 'mhm', 'shh', and 'tut tut'. These overlap with: (B) PAUSE PHENOMENA, including actual silence and 'er'. (C) NON-SEGMENTAL FEATURES - prosodic features - meaningful contrasts 'due to variations in the attributes of pitch, loudness, and duration' (94). How loud, how low or high, how fast, how rhythmical - as long as it makes a difference in the nature of the meaning of the 'message'. Paralinguistic features - everything else. A lot of work - and a lot of thinking - needs to be done here. Crystal's effort is valiant and careful, and should stimulate a lot of rethinking. The next chapter, (6), is an elaboration of (5) and provides some data - an 'illustration from religious language', in Crystal's system: why/how language and society are the same/related. After a complicated introductory page, in which he discusses 'three current emphases, which I loosely label empirical, methodological and theoretical' (96), he declaims against the concentration on syntactics to analyze paralanguage. He argues that 'most sociolinguistic distinctiveness in speech is phonological in character . .. primarily in its use of prosodic and paralinguistic features' (97). He uses examples from religious language for various reasons: the elaborated and ancient texts, the repetitiousness of ritual and its many categories about which we can ask if something is 'religious speech' or not, and its resonance all over the community - a kind of transcendence which enters into one's life, literature, and general language. Again Crystal places 'ownership' of sociolinguistics in the ears and mind of the analyst, not in the experience of the community: 'All sociolinguistic investigation commences by assuming on intuitive grounds that a particular category of language-situational co-variant exists: the subsequent analysis is then intended to verify this intuition' (98). The analyst begins as a kind of typologist, and 'recognizes' several distinct 'genres' or modalities (here, e.g., sermons, litanies, biblical readings, etc.). Within the same paragraph, the author turns himself from the 'intuitive identifier' to the active observer who actually checked out what he observed, with participants who were asked to verify his 'intuitions'. At some level, the author and this reviewer disagree about the nature of intuition vs. observation and who judges what is significantly meaningful. Here he begins as the pure metaphysician 434

REVIEWS

yet feels some need to check his categories. One is confused by what Crystal says, but senses that he ends up wanting to be some sort of social scientist, inferring from what he does. Four points emerge: (X)vocabulary,(2) syntax,(3) segmentalphonologydo not vary; and leaving (4) non-segmentalphonologywhere the action is. By observation (with some data), it is clear to Crystalthat these four modalitiesare distinct in ways that are not difficultto describe. For example, biblical readingis obvious to the observer, characterizedby a 'regularity of the speed and rhythm, the tendency of the intonationto follow the punctuation,the predictableoccurrence and length of pauses, . . .' (I03). That is, if people hear biblical readingas distinct (which they usually do), then we can figure out ways of describing it. These descriptive designations remain 'rough', and Crystal thinks various modalities may occur as several rather distinct systems. But this all seems correct and examinable in principle and 'indicates very clearly the interdependence of sociolinguistics with other aspects of linguistics' (104). Chapter 7 could be another version of the previous chapter, using poetry as an example, rather than religious language. The sheer size and variety of claims in metrical theory must be dealt with, however, and Crystal's ambitions take him on this excursion into Oz where descriptive or experienced reality is often beside the point. One senses that ultimate prestige for humanistsis located somewherein the arenaof literarycriticism,includingmetricaltheory,and 'drives' the linguist to prematurelyjoin this crowd, as if convincing 'them' of some linguistic truth would yield entry to 'the' intellectualclub. The chapter is a full 20 pages; the major points could be made in three or four. Crystal'sattackbegins by claiming that the general view of metrical theorists is that 'metre is held to be an abstraction,in some sense, and is not to be identified with performance'(o05). Obviously, the argument can rage forever, since it remains ungrounded in any possible common experience; one more arena for playing the 'Glass Bead Game'. The attempt to enter this game critically is beset with the problemcommonto convincing'believers';i.e. criticismis assumed to be coming from 'within' the belief system, thus is easy to dismiss merely in terms of its heterodox tone. An example is Crystal's definition of meter: 'The hierarchicsystem of continuous recurrentnon-segmental phonologicalequivalences which constitute the organizing principle of a poetic text' (xo7). This doesn't help much. He could have said that it is 'agreed' that certain events/ texts are poetry - what is distinctiveabout them?- what arethey? - how are they differentand distinct?- just as he did for religious language,where he successfully avoided this over-systematizedform of argumentation. More usefully, Crystalwonderswhy 'stress'has been emphasized,and 'intonation' discountedin metricaltheory (but then, why call it 'metrical'?). Apparently, it is little more than custom, but the extent of dogma within which poetry-asstress is embedded,speaksto its lackof groundednessin fact. In any case, Crystal 435

LANGUAGE

IN

SOCIETY

wants to reject most of the previousideas; citing and rejectingthem, at length. He (I looksfor 'some supportin the previousliterature' 14), and, of course,finds some. He presents some data with his analysis(120), claiming that: (i) all lines were coterminous with tone-unit boundaries(one sort of exception); (2) 8o0/ of all lines consisted of a single tone-unit; (3) within lines, points of prominencewere and usually pitch-contrastive,not loudness-contrastive; (4) the range of features needed to transcribethe contents made use of in the poetry readingswas much greaterthan that needed for prose. This satisfiesCrystalthat, 'in poetry we seem to be dealing with a distinct non-segmentalphonologicalsystem, and the range of distinctivenessis best describedby referenceto the unit line' (12I-2). This is probably some form of truth, but one wonders if (as the believers in metrical theory will say) the author's system drives him to this conclusion irrespective of the data or analysis. Crystalsets up a complex defensive barragein which his claims are embedded. They seem not to add to the force of his claims, which might be correct in spite of his attempts to qualify them. The fact is, that Crystal's approach is quite revolutionary at least in his mind. Qualifying what one is doing, to the present holders of (intellectual, literary . . .) power, will not affectthem and does not seem particularlyattractive to hi-, potential followers. In the author's concluding remarks to Chapter 7: 'I have tried to suggest a whole new orientation...' (124). Chapter 4 tries to tell us why. The concluding chapter (8), much like (6) and (7), is anotherarea where it is useful/necessaryto bring in paralinguisticthinking: to understandthe natureof language acquisition. Although this chapter is largely theme and variation, it has yields more clues to the author'sassumptiveorientations.Paralanguage been as neglected in languageacquisitionas it has been in adult speech and for much the same reasons. In addition, there are difficulties in 'obtaining natural and reliable samples of speech for analysis and of relating non-segmentalcharacteristics to other (and even less studied) communication modalities, such as the visual and tactile - a particularproblem for the study of young children, where kinesic and other cues regularlyneed to be taken into account for any complete interpretationof non-segmentalpatterns(I25). Since this reviewerbelieves this is all necessaryfor adultsas well as young children,it is interestingto wonderwhy Crystalmakes this area a special case in particularways. It apparentlyhas to do with the sort of belief that developmentof languagehas mostly to do with logic and increasingrationality,and depends for its underlying'theory' on particular definitionsof rationalityfor a developmentalscheme. (This also virtuallyeliminates the possibility of seriously examining languagein any cross-speciessense.) It also tends to lead away from the idea that children's speech is intrinsically interactionaland 'meaningful'and towardsome notion of 'levels' of development. How one comes to all of these seems to have a heavy influenceon what and how one studies as language acquisition. 436

REVIEWS

Examples: 'I shall be using the term vocalization. . . for which there is no evidence of language-specific contrastivity,i.e. the sound patternsare biologically controlled' (127). This would include 'crying', 'cooing', etc. To the extent that parents respond to cries and coos, they probably help 'shape' the meaning (at least the 'social meaning') of what cries and coos are about. That is, if parents respond to 'biologically controlled' sounds as if they are meaningful- and the infant uses them - then they are, for all intents and purposes, meaningful. And so we come to another form of argument about the nature of linguistics and where 'meaning'is located. Crystalhas probablymade deeper assumptionsabout the (individual) nature of human nature than will serve his purposes. In his scheme, the period prior to a child's 'evidencing phonologicalcontrastivity' is 'pre-linguistic' and not amenable to his analysis. How anyone comes to 'have' languagewill forever remaina mystery belonging to 'biology'. Again, one may detect another form of dualism - culture vs. nature- and what Crystal is proposing is a mere shift in boundaries. The author regards any division of this field as 'inevitably arbitrary',but 'conveniently'breaksit up into three areas: (i) 'early, "pre-scientific"investigaanalysisof vocalizationsin young infants'; and (3) 'studies tions'; (2) 'parametric of non-segmental patterningin older children' (130). is and Early research not well surveyed,except impressionistically suggestively. Here Crystal is driven to consider historical documents, especially Darwin's I877 paper on infant vocalization. Most thinkers impute some innate abilities to each infant which effectively 'cause' future language development. Some sort of emotional attribution is usually cited. Typical is Bridges (1932): 'out of an initial state of "excitement" there develops a distinction between delight, distress and excitement by three months; distress divides into fear, disgust and anger by six months, elation and affection . .. jealousy . .. joy' (132). Crystal is justly critical of this sort of description; and also criticizes the other sort of accountingwhich says that infants learn intonation patternsfirst and then come to segmental phonemes. While this appeals to most linguists as an accounting for aspects of language they consider to be someone else's problem, 'the main weakness in this work . . . is the absence of precise information about the characteristicsof the adult intonational stimulus' (I35). A term such as 'babytalk' is widely used as an explanatory device without any clear or accurate specificationof what 'baby-talk'might be. Crystal concludes that most previous work in this area is neither well-conceived nor well done. Recentanalysisof vocalization:Here Crystalis more optimistic, programmatically suggesting a variety of approachesto the problems of language acquisition. He believes they will be primarilyacoustic. (One wonders if childrens' ears are, even partially, tone analyzers; but one could study how they come to be. The Helmholtzian assumptions underlying spectrographicanalysis are not critically reviewed by Crystal!) He pays some attention to other sorts of approaches,and 437

LANGUAGE IN SOCIETY

their problems and advantages: ethological and social-behavioral,but strongly leans on acousticsolutions, presumablybecause that field is better developedand at least accurateand consistent in its own terms. The book ends with a fairly extensive review of essentially the entire field, in as it has focused on: Non-segmentalpatterning olderchildren.In effect, this is an essay on the state of the art, Crystalraising critically all of the problems as they have arisen under this rubric. His concluding paragraphtells us much: 'As Lewis (1936: 95) said: "The whole question of intonation in children's speech is ... extremely obscure." It is depressing, nearly forty years later, still to have to agree with him' (158). This reviewer, having spent considerable time and effort to enter into this book, found it a very worthwhile experience. The book represents a major criticism of linguistics at the level of what the subject matter properly'is'. In this sense, much of the presentationis a vehicle for, and a buttressing of, Crystal's arguments. To his great credit, he lays out a positive system, a 'new' way to do linguistics, which extends it more toward an examination of the human experience as a meaningful venture. While this reviewer is critical of this system for reasons cited, he appreciates the willingness of Crystal to subject his ideas to public criticism. They deserve to get a full, thoughtful hearing. Because Crystal is careful and thorough, he has covered the contemporary field fairly completely. There is a very extensive bibliography- some 30 pages; it is difficult to say what gaps there are, or if they are systematic. On the other hand, Crystal presents few examinations of the historicalphilosophicalbackgroundand underpinningsof his argumentation. The title, The Englishtone of voice, is gratuitous; it is a cover for this wellreasoned attempt to extend linguistics toward the wider arenas of meaning, of semiotics, of human communication ... It might be more appropriately entitled '(Human) tone-of-voice'. What Crystal says is that language occurs in real-social situations; what is linguistic is what makes a difference in interaction,irrespectiveof whether it is phonemic in any (earlier)structuralsense. While there has been an implicit, hidden theory of meaning in linguistic theory, the author attempts to make explicit and to broaden the kind of meaning (theories) in terms of which our behaviortakes place. Crystalwants to extend this to much or most of the verbal behavior which occurs in spontaneous conversation.He is unwilling to extend it to non-verbal behavior- at least so far - thus casting his lot with linguistics rather than with semiotics.
REFERENCES Bridges, K. M. R.
(1932).

Emotional development in early infancy. Child Development

324-41.

438

REVIEWS

Prosodic systems and intonation in English. Cambridge: Cambridge Crystal, D. (I969). University Press. Crystal, D. & Quirk, R. (I964). Systems of prosodic and paralinguistic features in English. The Hague: Mouton. Reviewed by HARvEr SARLES Department of Anthropology University of Minnesota Minneapolis, Minnesota 55455 USA

(Received

19

July 1978)

STARKEY DUNCAN,JR. & DONALD FISKE,Face-to-face interaction: W. Research, methods, and theory. Hillsdale, New Jersey: Lawrence Erlbaum Associates (distributor Halsted (Wiley)), New York, 1977. Pp. 36I. The face-to-face interaction examined in this work is restricted to conversation, though it includes detailed analysis of non-vocal as well as vocal behavior of participants. The core of the book consists of studies of videotaped, two-person conversations from two very different perspectives: first, an 'external variable study' correlating selected actions of each participant with other acts that he or she performed, actions of the partner, and a range of measures obtained from self-reports; secondly, a 'structural' study investigating sequences of actions as an organized system with reference to the issue of how turns are exchanged in conversation. There is, in addition, a preliminary methodological section in which the authors argue against attempts to control variables by manipulating confederates' actions and in favor of an empirically based, naturalistic research strategy. A concluding section proposes a 'metatheory' focusing on the study of conventions, situations, and interaction strategies. One of the very strong points of this work is the comparison its two studies provide of alternative approaches to the study of face-to-face interaction. External variable studies are, of course, the mainstay of much research in the contemporary social sciences, including some approaches to sociolinguistics. The study reported here is more comprehensive than most, including almost 50 variables and 88 subjects and leads to a variety of substantive findings (for example, very little correlation between self-reports and observed behavior). The authors, however, find that it tells them far less about the structure, organization and details of the phenomena being investigated than the structural study, stating explicitly (123): 'Our substantive findings from our studies of correlates of acts in interaction do not impress us.' They argue (13I-2) that among the major failings of a study of this type is the fact that it separates acts from the specific acts preceding them, thus treating context as a global variable, when, in fact, the most relevant context for the production of a particular action would seem to be the immediate interactive one and especially the acts that the other participant has just performed. In an attempt to deal with such issues, the structural study took as its point 439