LINGUIST List 19.3804

Fri Dec 12 2008

Review: Semantics: Lee, Gordon & Büring (2007)

Editor for this issue: Randall Eggert <>

        1.    Philip Davis, Topic and Focus

Message 1: Topic and Focus
Date: 12-Dec-2008
From: Philip Davis <>
Subject: Topic and Focus
E-mail this message to a friend

Discuss this message

Announced at
EDITORS: Lee, Chungmin, Matthew Gordon & Daniel BüringTITLE: Topic and FocusSUBTITLE: Cross-Linguistic Perspectives on Meaning and IntonationSERIES: Studies in Linguistics and Philosophy 82PUBLISHER: SpringerYEAR: 2007

Philip W. Davis, Rice University

SUMMARYThis volume contains fourteen chapters that are the product of a workshop heldin 2001 during the LSA Summer Institute at UC Santa Barbara. ''The workshop wasdesigned to lay the groundwork for collaborative efforts between linguistsdevoted to the study of meaning and linguists engaged in the quantitative studyof intonation'' (vii). Eleven of the chapters are descriptive studies of Topic orFocus. Three (Gil, Steedman, von Heusinger) are concerned with other semanticsof intonation. Twelve languages constitute the empirical base: Basque(Elordieta) Polish (Eschenberg), Riau Indonesian (Gil), Chickasaw (Gordon),English (Gussenhoven; Hedberg & Sosa; Krifka; Steedman; von Heusinger), Dutch &Italian (Krahmer & Swerts), Korean (Lee), Japanese (Nakanishi), Taiwanese (Pan),Bengali (Selkirk), and German (von Heusinger). Focus receives the mostattention. It is the subject of chapters by Elordieta, Eschenberg, Gordon, Gil,Gussenhoven, Hedberg & Sosa, Krahmer & Swerts, Krifka, Pan, and Selkirk. Topicis addressed by only three authors, Hedberg & Sosa, Lee, and Nakanishi. Steedmanand von Heusinger address situational semantics more broadly. Each chapter hasits own bibliography, and there is no index.

In the review that follows, the chapters are discussed in their printed order inthe collection. The first paragraph gives a straightforward description of whatI think the chapter is about. The second paragraph, if there is one, containsmore evaluative remarks about the chapter. At the end of the review, I addevaluative comments about the book as a whole.

CHAPTERSGorka Elordieta. ''Constraints on Intonational Prominence of FocalizedConstituents''. Elordieta discusses one dialect of a variety of Northern BizkaianBasque, Lekeitio Basque (LB). The language is S IO O V, with the expression ofFocus placed in immediate preverbal position and accompanied by ''intonationalprominence'' (1) or ''main prominence'' (3). ''Prominence is realized as H*+L pitchaccent'' (6). LB distinguishes in this way ''neutral declarative sentences'' (3) or''broad focus'' (11) from utterances with ''narrow focus'' (9) and from those with''corrective focus'' (15). Elordieta is concerned more specifically with examplesin which the preverbal constituent is compound, containing a genitive possessorfollowed by a possessed. In that context, a distinction between accented andunaccented words (5) is necessary since only accented words may carry prosodicprominence. An unaccented word may acquire a ''derived accent'' (5, 12) by beingthe second of the possessor + possessed pair, but not the first. If bothpossessor and possessed are accented lexical items, either may be differentiallynarrowly focused. But there arises an asymmetry in the expression of Focus inthe possessor + possessed pair when the possessor is unaccented. The inherentlyunaccented possessor does not have derived accent and cannot therefore be moreprosodically prominent. This asymmetry is absent when Corrective Focus isexpressed. The pattern of focal expression is further modulated by variationamong speakers.

Ardis Eschenberg. ''Polish Narrow Focus Constructions''. Eschenberg describes theoccurrence of Focus with the S and O constituents of Polish. Assuming that thelanguage has a neutral order of SVO (23), Narrow Focus is marked by ''prosodicprominence'' (30) and order. If the S or the O occur in their neutral positionsand carry the prominence, the result is an expression of Informational Focus.The non-neutral orders of OV or VS accompanied by intonational prominence on theO or the S yield Identificational Focus (31, 39). S and O constituents differ intheir co-occurrence with 'even' and 'also' and intonational prominence. An Omust appear only in the neutral position of Informational Focus, while the S maybe used in both the SV and the VS positions. In the VS position, the Focus is''presentative'' (33). Eschenberg's two spectrograms (26) show that Polish - likeLekeitio Basque - prosodically distinguishes a Corrective Focus, his ''correctionparadigms''. From the cited examples, immediate preverbal position and immediatepostverbal position seem to be the ones involved in noncanonical Focus. AlthoughEschenberg asserts that it is more accurately sentence final position that isrelevant, i.e. Polish has a VOS order (33), but not a VSO, the chapter containsno example to show this. All examples with a postverbal transitive S have the Oelided.

Since several of Eschenberg's examples (PIOTR spiewal 'Péter sang' [25], SpiewalPIOTR 'Péter sang' [25], and SPIEWAL Piotr 'Peter sáng' [29]) show that an S mayappear stressed or unstressed both before and after a V, it may be that wordorder is independent from the prosodic prominence of Focus, representing aseparate grammar and a separate semantics. O's do not interact with word orderin the same way as S's. There may then be at least three components to thepatterns of Polish Narrow Focus: the syntax and semantics of word order itself,the intonation and semantics of Focus, and the semantic makeup of the S and Ofunctions. In addition to the semantics of 'agent' and 'undergoer', the S andthe O functions seem to have additional semantic coloring that associates themwith Focus in contrasting ways.

David Gil. ''Intonation and Thematic Roles in Riau Indonesian''. The ''mainconcern'' of Gil's chapter is to disprove ''the purported correlation betweenintonation and thematic roles'' (57) in Riau Indonesian. He does this in thefollowing way. The four basic Riau Indonesian intonation contours (57) are shownto occur with each of the four basic sentence patterns (58): Actor precedesactivity, Undergoer precedes activity, Actor follows activity, and Undergoerfollows activity. Then it is demonstrated that none of the cooccurrences ofintonation contours and sentence patterns has a significantassociation/disassociation. Because of the absence of any correlation betweenintonation patterns and the presence of a role, the two are independent andintonation does not signal role. Riau Indonesian sentence patterns differ (fromthe Eurocentric perspective) in the absence of a contrast between actor andundergoer (63), either preceding or following the activity.

Except for a brief excursus into the use of intonation patterns to signal Focus(54-56), this chapter appears to have only slight connection with the theme ofthe book as a whole. It is, however, interesting on other grounds. Gil claimsthat Riau Indonesian fails to distinguish roles in any fashion, grammatically orsemantically, and thus organizes its propositions following principles otherthan roles and voice.

Matthew Gordon. ''The Intonational Realization of Constrastive Focus inChickasaw''. Gordon ''examines the prosodic realization of sentences involving thecontrastive focus on subjects and verbs [actually objects, PWD]'' (71).Contrastive Focus is marked by -akot for subjects and by -akõ: for objects.Fundamental frequency and duration are the two prosodies examined. Differencesin both frequency and duration are present and associated with ContrastiveFocus; but since neither is contrastive, their use is variable across categoriesand across speakers.

The technique which yielded these results relied on Chickasaw speakers' abilityto respond to an English stimulus: ''Focus was elicited by offering Englishtranslations emphasizing the focused element'' (71). It is not clear how, orwhether, the English stimulus in fact reflects Chickasaw semantics, especially''since the precise semantic conditions that give rise to contrastive focus [inChickasaw] are not completely understood'' (71).

Carlos Gussenhoven. ''Types of Focus in English''. Gusshoven's chapter on Englishconsists of two parts. The first contends that ''the way pitch accents expressinformation structure in English is subject to structural constraints'' (83),paraphrased either as ''the 'focus-to-accent' relation ... [is] indirect, andmediated by the linguistic structure'' (85) or as ''the relation between the pitchaccent and the focus is mediated through the predicate-argument structure of thesentence'' ( 97). The Stress Accent Assignment Rule (87) is a concrete embodimentof this assertion and demonstrating its working is support both for the SAAR andfor the supposition that has created it. The second portion of the chapter listsand illustrates seven types of Focus: presentational, corrective,counterpresuppositional, definitional, contingency, reactivating, andidentificational.

There is so much to react to here that it is not possible to do so without beingarbitrary. Certainly Focus interacts with its semantic environment, butattempting to understand the semantics of Focus and the nature of itsinteraction with its semantic context(s) might constitute a reasonablealternative approach to more mechanical ones that use expressions such as''causing the predicate to be accented'' (87) and ''an indirect object ... licensesthe unaccented predicate'' (88-89). Such an approach might also provide aperspective on the types of Focus. Are there only seven? Why not eight or six?How are they interrelated? Cf. the Conclusion below.

Nancy Hedberg & Juan M. Sosa. ''The Prosody of Topic and Focus in SpontaneousEnglish Dialogue''. Hedberg & Sosa present an analysis of spoken English from theperspective of information structure and intonation (101 et passim).''Information structure'' here means five categories: contrastive focus, plainfocus, contrastive topic, unratified topic, and ratified topic (101-102).''Intonation'' is represented in terms of the tones and break indices notation(ToBI Labelling). The strategy was for the first author to examine a writtentranscript of the text to be analyzed (a television broadcast), using a priorispecifications of the five information structure categories and marking the textwhen they were thought to be present. The next step was for the first authorthen to listen to the videotape of the text to ''confirm these codings'' (101).The second author examined a select portion of the text and assigned the ToBILabelling. The ''major goal'' (108) was to examine several ''hypotheses'' about therelation between the categories of information structure and their associationwith specific ToBI labelled intonations. Any connection between the two appearsto be partial. The results of the study are described in terms of a ''best fit''(111), categories that ''are only sometimes marked'' (114), uses of intonationthat ''contrary to the predictions in the literature'' (115), and an hypothesisthat ''is not borne out by the data'' (118). The one mostly positive discovery isthat ''Except for ratified topics, which tended to be unaccented, most phrases ineach information structure category are marked H* [a peak accent, PWD]'' (112).

I find the technique puzzling. Why would one assume that any specimen of anylanguage communicated some meaning without simultaneously noting the portion ofthe utterance that gave expression to that meaning? The interesting question inthe analysis is ''What did the first author in fact hear that confirmed theinitial codings?'' If it were those ToBI Labellings, the analysis could neverhave been completed.

Emiel Krahmer & Marc Swerts. ''Perceiving Focus''. Dutch employs pitch accent toindicate the locus of content being focused. Italian does not. Krahmer & Swertsdevise an experiment to demonstrate that speakers of Dutch can perceivealternative placements of pitch on an adjective + noun sequence while speakersof Italian ''fail completely'' (135) in an analogous task.

If pitch accent marks Focus in a language, then its speakers can by and largehear it in an experimental context, and if it is not so used, speakers gain nomeaning from it, and they fail in the same task. They do not hear it. Theauthors seem to have discovered the Phonemic Principle (Swadesh 1934).

Manfred Krifka. ''The Semantics of Questions and the Focusation of Answers''.Krifka considers the English expression of alternative questions, multipleconstituent questions, and the Focus patterns of answers to constituentquestions. He is concerned to demonstrate that ''Alternative Semantics does notpredict the correct patterns of answer focus ... [and that] The StructuredMeaning theory, on the other hand, does not have these problems'' (139). Krifkaconcludes that ''it appears that the careful consideration of focus in answers toconstituent questions argues against the alternative semantics account, and forthe structured meaning account, of questions and answers'' (150).

The proof of Krifka's thesis relies upon a formal logic, the notation of whichthe reader must be familiar with in order to gain access to the content of thechapter.

Chungmin Lee. ''Contrastive (Predicate) Topic, Intonation and Scalar Meanings''.Lee demonstrates the existence of a Contrastive Topic (CT) in Korean (157),distinguished from Non-Contrastive Topic by intonation (153). It is elicited bycontexts such as this (157): ''After hearing that Inho didn't come, regarding hisfriend Yengswu'' the speaker utters ''Yengswu-nun w-ass-e'' (Yengswu-CTcome-PAST-DEC) 'YengswuCT came'. ''The crucial requirement of CT is thatpotential Topic of sum must precede or be assumed to precede it'' (158). TheKorean CT has a specific relation to scalar implicatures: ''A typical CT with anappropriate contour evokes a scalar implicature conventionally by default ...''(161).

Kimiko Nakanishi. ''Prosody and Scope Interpretations of the Topic Marker wa inJapanese''. Like Korean, Japanese has a contrast between a Contrastive Topic anda Non-Contrastive one (Nakanishi's ''thematic wa'' [177]) distinguished byintonation pattern (181). Nakanishi demonstrates the contrast and thendemonstrates a corresponding contrast in scope when the Topic is minna'everyone' and the predicate is negated: Minna-wa ne-nakat-ta (everyone-TOPsleep-NEG-PAST) 'Everyone didn't sleep'. When minna-wa has the intonation of aNon-Contrastive wa, the sense of ''It is the case that everyone did not sleep'.No one slept. When the intonation is contrastive, the sense is 'It is the casethat not everyone slept'. There is someone who didn't sleep (182-183).

Nakanishi (179) illustrates the sense of 'contrastive' using the utterance''Naoya-wa nonbiri-si-teiru ga Maria-wa nonbiri-si-tei-nai'' (Naoya-TOP relaxdo-PROG but Maria-TOP relax-do-PROG-NEG) 'Naoya is relaxing, but Maria is notrelaxing', with Contrastive intonation on Naoya-wa. But 'contrastive' may be toonarrow a semantic interpretation of this intonation. ''Naoya-wa nonbiri si-teiru''(Naoya-TOP relax-do-PROG) 'Naoya is relaxing', with contrastive intonation, canalso be pronounced as a retort to the assertion ''Naoya-wa nonbiri-si-tei-nai'''Noaya is not relaxing', with Non-Contrastive intonation, to yield a sense of'Naoya is too relaxing [contrary to what you said]'. (Personal communicationfrom a native speaker of Japanese, which I hope I have not misunderstood.) Sinceboth the assertion and the retort contain the same Topic, Naoya, the meaning'contrastive' requires some modification, perhaps something in the direction of'insistence'. Kuroda (2005:7) shows that ''thematic wa'' may carry the meaning ofFocus in that it can answer wh- questions. It might be interesting to considerthe possibility that contrastive intonation with wa is signaling the presence ofsome variety of Focus on that constituent. Kuroda's (2005:26-28) contrast of''asserting'' and ''affirming'' (''wa sentences assert while non-wa sentencesaffirm'') may be relevant here if the contrast describes a dimension in place ofa dichotomy. Lee (153) notes an analogous debate in the literature concerningContrastive Topic in Korean.

Ho-Hsein Pan. ''Focus and Taiwanese Unchecked Tones''. Taiwanese has seven lexicaltonal contrasts, which have morphophonemic variants depending upon where theyoccur in a ''tone group''. Narrow Focus is manifest suprasegmentally in Taiwanese,and Pan describes its allomorphy in the context of tone morphophonemics and alsoin the context of the syntactic position of Focus in the utterance. Panconsiders the positions of Focus in a short SVO utterance. Focus is realized byan intonational combination of duration, range in fundamental frequency (F0),and in mean F0 values.

Elisabeth Selkirk. ''Bengali Intonation Revisited: An Optimality TheoreticAnalysis in which Focus Stress Prominence Drives Focus Phrasing''. Bengali is anSOV language that distinguishes intonationally between (i) a Neutral or BroadFocus declarative expression, (ii) the same as (i), but a yes-no question, (iii)a declarative expression that contains an Informational Focus either on the V ora constituent preceding, and (iv) the same as (iii) but a yes-no question. TheInformational Focus is a H+L+H contour that covers the focused constituent(233). The issue of the chapter is the presence of the Grenzsignalende second Hof Focus. ''Focus Prominence Theory ... predicts a phonological phrase edge atonly one edge of a focus constituent, the edge where the focus prominence islocated ... [but] a focus constituent in Bengali is flanked by phonologicalphrase edges ...'' (216), i.e. the two H's. Selkirk solves the contradiction with''constraints and rankings'' (240).

The Bengali equivalents of 'Did I give money for the king's pictures?'' withNeutral Focus (227) and the declarative 'I gave money for the king's pictures'with Focus on the V (233) have these intonational contours, respectively: L*H L*H L* [HL]QUES and L*Hp L* Hp L* H[L]. Removing the nonphonetic parts of thenotations, there is one intonation, L*H L* H L* HL, and there appears to be asingle Bengali sentence with two glosses. Selkirk does not discuss this homophony.

Mark Steedman. ''Information-Structural Semantics for English Intonation''.Steedman discusses English pitch accents, e.g. L+H* and H* (248-250), incombination with ''intonational boundaries'' (247), e.g. LL% and LH% (250 251).The meaningful contrasts of the pitch accents are attributed to ''informationstructure'' (245), i.e., '''not given' information'' (246) and ''theme'' & ''rheme''(246), and to ''contentiousness'' (245), i.e. whether ''mutually agreed'' (246) ornot. The contrasts in intonational boundaries ''distinguish the speaker or thehearer as responsible for, or ... committed to, the corresponding informationunit'' (247). The resulting patterns are expressed in terms of Steedman'sCombinatory Categorial Grammar (255 259).

The interactional, situated meanings which Steedman treats are always difficult.They themselves can be contentious, and for that reason, they are interesting.In the nondiscrete world of pitch and intonation, it is not surprising that''trained ToBI annotators show quite low inter-annotator reliability'',particularly in distinguishing between H* and L+H* (259). This reprises aconcern I had above about Hedberg & Sosa's chapter. Situated expectation as wellas phonetic skill (professionally acquired or native) will affect what we thinkwe hear. More than once, I have thought, ''I didn't mean to say what you thoughtyou heard.'' Explaining to your spouse that you lost control of the intonationalcontour is no excuse. You said it, you meant it. One hears meaning, not sound.

Klaus von Heusinger. ''Discourse Structure and Intonational Phrasing''. Settingaside the ''discourse functions of pitch accents and boundary tones'' (that werethe focus of Steedman's chapter) and drawing examples from German and English,von Heusinger ''argue[s] that intonational phrasing determines minimal discourseunits which serve as the building blocks in a discourse representation'' (265). Adiscourse ''consists of sentences that are related to each other by relations,such as causation, explanation, coherence, elaboration, continuation ..., whichcan then be represented in a tree ...'' (270, 283). von Heusinger ''argue[s] thatthe semantics of intonational phrasing can best be accounted for in terms ofdiscourse units ... defined by their function to serve as arguments in discourserelations'' (266, 283). von Heusinger elaborates briefly five such discourserelations: non-restrictive modification, backgrounding, enumeration,topicalization, and frame-setting (285ff.) The result is expressed in terms ofDiscourse Representation Theory (266 et passim).

CONCLUSIONThere is no such thing as an objective evaluation. The fairest I can be is toadmit to the persistence of bias, explain briefly what those biases are, andthen try to be consistent within them. My belief has been that language existsbecause it means and that the form of language is secondary and exists onlybecause it has meaning. The possible formal resources to express meanings arecontrasts in linear sequences, paradigmatic substitutions, and suprasegmentalcontrasts (essentially, Bloomfield's taxemes). The complexity of language liesin the meaning and not in the form. Certainly, any ''explanations'' of languagefollow from understanding its meaning.

The title ''Topic and Focus'' is what drew me to this book. I was even moreexcited to see a subtitle ''Cross-linguistic Perspectives on Meaning ...'' andthen ''and Intonation''. Intonation was less attractive, but still OK. Thechapters are descriptive studies of specific languages from this perspective,and the promise is that the reader will find unknown ways these languagesconstruct and organize the semantics of Topic and Focus ... in so far as they doit with intonation.

So what is there new about the semantics of Topic and Focus?

There is an impressive number of Focuses cited. I counted at least 16, and I mayhave missed some. While we find familiar, traditional terms used, others aremore special. These are the Focuses I have recorded: ''broad'' (e.g. Gordon, 71 &73; Krahmer & Swerts, 130-131, Pan 199) versus ''narrow'' (e.g. Elordieta, 9 & 11;Eschenberg; Gordon, 71 & 73; Krahmer & Swerts, 130-131; Pan, 199),''informational'' (e.g. Eschenberg, 31) versus ''identificational'' (e.g.Eschenberg, 31; Gussenhoven, 96); ''corrective'' (e.g. Elordieta, 14-15;Eschenberg, 26; Gussenhoven, 91); ''contrastive'' (e.g. Gordon; Hedberg & Sosa,111; Selkirk, 217, 221); ''presentative'' (e.g. Eschenberg, 33), ''presentational''(e.g. Gussenhoven, 91); ''counterpresupposition'' (e.g. Gusshoven, 92),''definitional'' (e.g. Gussenhoven, 92); ''contingency'' (e.g. Gussenhoven, 94);''reactivating'' (Gussenhoven, 95); ''plain'' (e.g. Hedberg & Sosa, 112); ''neutral''(e.g. Selkirk, 217), and ''big'' versus ''small'' (Selkirk, 220). Some of thevariety appears to be just a matter of labels. ''Broad'' Focus seems to be thesame as ''plain'' and ''neutral''. Answers to wh-questions are ''corrective'' forElordieta (14), ''narrow'' (either ''identificational or ''informational'') forEschenberg, ''presentational'' for Gussenhoven (91), ''narrow'' for Pan (199), and''contrastive'' for Selkirk (221). Clearly, authors will use the terms a bitdifferently. That is not really a complaint, but the proliferation of types maybe. Clark & Marshall (1981.22 23) enumerate ''eight major uses of the [English]definite article'': the anaphoric use, the visible situation use, the immediatesituation use, the larger situation use based on specific knowledge, the largersituation use based on general knowledge, the associative anaphoric use, theunavailable use, and the unexplanatory modifier use. The list omits the genericuse. The question is where does the list end? Is there a tenth use somewherewaiting to be discovered? Such a circumstance indicates to me that a fundamentalunderstanding is missing, and I sense the same in the large number of Focusescited in these chapters.

Consider briefly Identificational Focus. Kiss (1998.245) has associated it with''exhaustive identification.'' Identificational Focus has been proposed forEnglish (Kiss 1998.256-260 et passim, Eschenberg, 31, Gussenhoven, 96) andlocated in cleft sentences. I believe that in English, Informational Focus doesnot exist as such. It is a contextual variant of whatever English Focus isotherwise. Its establishment as a distinct type of results from a failure toseparate the semantic contribution of the pitch accent from the semanticcontribution of the morphosyntax of the clefting. Cp. ''It was the dark thatscared him'', with pitch accent alternately on ''dark'' and on ''scared'', or even''was''. And if the pitch accent is on dark, then the following ''that scared him''may have a lower level intonational contour to its end or a non-level contourwith a slightly higher pitch and prominence on scared. Intonational matters seemvery distinct from the morphosyntax. and within the morphosyntax of the cleft,there is a further contrast between the choice of it as subject versus someother pronoun, e.g. ''It's the plumber'' versus ''He's the plumber'' in response to''Who's that in the kitchen?'' (Declerck 1983). Each of which may be followed by''... that you called yesterday''. Only ''It's the plumber'' is a felicitous answerto ''Who did you say is coming this morning?'' Identificational Focus is theoutcome of choosing copular morphosyntax with a subject ''it'', choosing (or not)to follow the copular complement with a dependent clause (and choosing anintonational contour for that clause), and then placing the pitch accent ofFocus. Identificational Focus appears when all those semantics come together. InEnglish, Identificational Focus is at best an ''allo-Focus''. (Cf. also mycomments above on Eschenberg's description of Polish Narrow Focus.) This is notto say that a distinctive Identificational Focus is absent from all languages.

Except for descriptive information about specific languages, I think a broaderunderstanding of the semantics of Focus is little advanced by this book.

Topic is discussed in English, Korean, and Japanese. In their paper on English,Hedberg & Sosa employ a ''ratified'' Topic, an ''unratified'' one, and a''contrastive'' one (102 et passim). Hedberg & Sosa's primary goal is to identifythe Topics by their intonational marking (101, 112, 114), but I find nodiscussion of the semantics. Ultimately, the Contrastive Topic and theUnratified Topic are (apparently) dissolved and referred to as ''ContrastiveFocus'' (111). Lee relies on the common criterion of 'about' in discussing Topic;''It is something talked about by the Comment ...'' (Lee, 132). He extends'aboutness' to the Korean Contrastive Topic (154-155): ''... a CT is 'about' agiven part in the previous discourse and locally 'about' the rest of the CTutterance. Hence it is topical.'' (I think I understand what this says, but itseems to be expressed backwards; the CT is Topic not because it is 'about'something, but because something is 'about' it.)Nakanishi accepts wa as a marker of Japanese Topic and also Kuno's (1973)characterization of Topic as ''anaphoric or generic'' (187). 'Aboutness' is notinvoked for Japanese. Each author spends most their effort on the semantics ofContrastive Topic, either in terms of scalar implicature (Lee) or scope (Nakanishi).

The ''quantitative study of intonation'', which the editors consider to be one ofhe defining components of the book, is manifest primarily in the use ofspectrograms (Elordieta, Eschenberg, Gordon, Hedberg & Sosa, Lee, Nakanishi, andPan). Pan includes several tables detailing measurements of differing vowellengths. ToBI Labeling has recently become prominent in the notation of Englishintonation. The basic publication appears to be Beckman & Ayers Elam (1997). Itis accessible through this url: Some of the ToBI publications warn thatthe labeling is not a substitution for a phonetic notation. It is a broadphonetic system specific to English, and there will be as many ToBI labeling asthere are language varieties. In addition to English, it used in the chapters onBasque, Bengali, and German, but there is no comment about any language specificadaptation. This is not a review of ToBI labeling. Suffice it to say I miss Pike.

In sum, I learned something about the languages described, but much less than Ihad hoped about Topic and Focus. I was not greatly entertained. _Topic andFocus_ is no _Subject and Topic_.

REFERENCESBeckman, Mary E. & Gayle Ayers Elam. 1997. ''Guidelines for ToBI Labelling.Version 3.'' Columbus: The Ohio State University Research Foundation.

Clark, Herbert H. & Catherine R. Marshall. 1981. ''Definite Reference and MutualKnowledge.'' In _Elements of Discourse Understanding_, ed. by Aravind Joshi,Bonnie L. Webber, & Ivan A. Sag. Cambridge: Cambridge University Press.

Declerck, Renaat. 1983. '''It is Mr. Y' or ''He is Mr. Y'?'' _Lingua_ 59.209 246.

Kiss, Katalin E. 1998. ''Identification Focus versus Information Focus.''_Language_ 74.245-273.

Kuno, Susumu. 1973. _The Structure of the Japanese Language_. Cambridge: The MITPress.

Kuroda, S.-Y. 2005. ''Focusing on the Matter of Topic: A study of wa and ga inJapanese.'' _Journal of East Asian Linguistics_ 14.1-58.

Swadesh, Morris. 1934. ''The Phonemic Principle.'' _Language_ 10.117-129.

ABOUT THE REVIEWERPhilip W. Davis is Emeritus Professor of Linguistics, Rice University. Hisinterests have been language description and syntax and semantics.