Date: Mon, 12 Dec 2005 14:24:08 +1100 From: Andrea Schalley Subject: Semantics: Critical Concepts in Linguistics, Vol. IV-VI
EDITOR: Gutiérrez-Rexach, Javier TITLE: Semantics SUBTITLE: Critical Concepts in Linguistics SERIES: Critical Concepts in Linguistics PUBLISHER: Routledge (Taylor and Francis) YEAR: 2003
Andrea C. Schalley, School of Languages, Cultures and Linguistics & Language and Cognition Research Centre, University of New England (Australia)
[This is the second part of a two part review, covering Volumes IV - VI, and concludes with an evaluation of the work as a whole. -- Eds.]
VOLUME IV - THE SEMANTICS OF PREDICATES AND INFLECTION [ix+461pp.]
After having dealt with the semantics of noun phrases in Volumes II and III, Volume IV is concerned with what is called the 'semantics of predicates and inflection'. It contains 19 papers. The volume is again divided into two parts, Part A about 'Events, aspect, and thematic roles' (8 contributions) and Part B about 'Tense and modality' (11 contributions).
Part A - 'Events, aspect, and thematic roles'
Chapter 47 is a reprint of Zeno Vendler's ground-breaking and still highly influential article ''Verbs and times'' (pp. 7-22; Source: The Philosophical Review 56, 1957, 143-160). Vendler distinguishes four eventity types (or aspectual classes - eventity is a rather recent term for 'events and similar entities', cf. Zaefferer 2002): states, activities, accomplishments, and achievements. This classification is based on the observation that the ''use of a verb may also suggest the particular way in which that verb presupposes and involves the notion of time'' (p. 7), and thus is based on different time schemata. Vendler presents several tests that allow us to distinguish the predicates belonging to each class, although he already notes that there are ''verbs that call for two or more time schemata'' (p. 7) and are thus not unambiguously assignable to a class.
Terence Parsons' excellent contribution in Chapter 48, ''Underlying events in the logical analysis of English'' (pp. 23-58; Source: Ernest LePore (ed.), Actions and Events: Perspectives on the Philosophy of Donald Davidson; Oxford: Blackwell 1985, 235-267), is a reaction to Davidson's philosophy (who captured eventities as arguments in the logical form). He presents a Neo-Davidsonian approach (in which participants of eventities and semantic roles are represented separately). The purpose of his paper is to explore the idea that many sentences of English can be assigned logical forms that make reference to or quantify over events, states, and processes. Specifically, he deals with aspectual classes, perceptual verbs, adverbials of Manner, Location, Instrument, Direction, and Motion, and with the individuation of events.
Emmon Bach, in Chapter 49 ''The algebra of events'' (pp. 59-70; Source: Linguistics and Philosophy 9, 1986, 5-16), discusses parallels between the aspectual classification of verbal expressions and the mass-count distinction in the nominal system. Drawing heavily on Link 1983 (Chapter 41), he extends the algebraic treatment to the domain of what he calls 'eventualities' (roughly, what we have referred to as eventities), and explores analogies between events (a specific eventity or eventuality type) and plural individuals, and between bounded processes and portions of matter.
In Chapter 50, ''Aspectual asymmetry and quantification'' (pp. 71-102; Source: Veronika Ehrich and Heinz Vater (eds.), Temporalsemantik; Tübingen: Niemeyer 1988, 220-259), Henk Verkuyl argues that aspect is not determined by the verb and its intrinsic characteristics but is compositional in nature. More specifically, he elaborates the idea that aspect formation is asymmetrical in that the verb and its direct object noun phrase constitute verb phrase aspect, which has to be distinguished from the sentential aspect it yields in combination with the subject noun phrase. Using the verbal feature [+/- ADD TO] (a lexical feature applying to verbs and roughly distinguishing progress/dynamicity vs. stativity) and the nominal feature [+/- SQA] ('specified quantity of A'), he develops a system of composition of aspect. Also, Verkuyl interprets the verbal feature model-theoretically in order to account for the semantic interpretation of the verb phrase.
James Pustejovsky proposes an 'event semantics' in Chapter 51, ''The geometry of events'' (pp. 103-127; Source: Carol Tenny (ed.), Studies in Generative Approaches to Aspect. MIT Lexicon Project Working Papers 24; Cambridge, MA: MIT, Center for Cognitive Science 1988, 19-39), in which the topology of the event defines the aspectual classification. The focus is not on aspect of verb phrase or sentence, but on aspect as determined by subeventual structure of the verb. He defines a calculus of aspect, with verbs being represented as a sequence of events and states. As a result of this, thematic relations are merely a derivative notion. Pustejovsky also examines how this approach may provide a solution to adverbial modification.
Chapter 52, ''Thematic roles and their role in semantic interpretation'' (pp. 128-147; Source: Linguistics 22, 1984, 259-279) by Greg Carlson, is the first (alas, not convincing) paper of three papers in this part (and overall collection) concerned with thematic roles. Typically, thematic roles are seen as superfluous elements in model-theoretic semantic interpretations. Carlson suggests an alternative system of model-theoretic semantic interpretation which precisely requires the type of information thematic roles provide. He does this by including eventities as primitives in his system (rather than as derived elements of the model), and argues that thematic roles have an intermediate status between syntax and semantics.
Malka Rappaport and Beth Levin examine the notion of theta-roles or thematic roles in grammatical theory in Chapter 53, ''What to do with Theta-roles'' (pp. 148-176; Source: Wendy Wilkins (ed.), Syntax and Semantics 21, Thematic Relations; New York: Academic Press 1988, 7-36). They attempt to reconcile two general approaches to the use of thematic roles in the generative literature: on the one hand, thematic roles have been incorporated in numerous grammatical rules and principles; on the other hand, they have been criticised, leading some to conclude that they have no grammatical import and no status in syntactic theory. Rappaport and Levin claim that those two uses are a reflection of two distinct lexical representations. The first use is relevant to lexical syntactic representation, the second to lexical semantic representation, which leads them to conclude that theta- roles are not primitive at any of those two levels. Note that the references list is missing completely (and not just about half of it, as in Chapter 8).
Although David Dowty's innovative contribution ''Thematic proto-roles and argument selection'' (pp. 177-208; Source: Language 67, 1991, secs. 4-8, 560-582) is a journal article, Chapter 54 does not reprint it in full (but only secs. 4-8). One of the problems Dowty identifies with thematic roles is that we lack a principled way to decide what kind of data motivates a thematic role type. Therefore, he reduces thematic roles to proto-roles. He conceives of thematic roles as prototypes, ''because role types are simply not discrete categories at all, but rather are cluster concepts'' (p. 189). His conclusion is that we need only two prototypical role notions: Proto-Agent and Proto- Patient. Each of these two roles is associated with different entailments. Contributing properties of the Proto-Agent are volitional involvement, sentience, causation of an event or change, movement, and potentially independent existence. The contributing properties of the Proto-Patient are undergoing a change of state, incremental theme, causally affectedness, and potentially an existence which is not independent. On the basis of these entailments, Dowty introduces specified ways on how to attribute which proto-role to which argument.
Part B - 'Tense and modality'
Chapter 55, Hans Reichenbach's ''The tenses of verbs'' (pp. 211-220; Source: Elements of Symbolic Logic; New York: Macmillan 1947, 287- 298), takes us back in time to the 1940s. The extract reprinted here proposes an approach in which natural language tenses are interpreted with respect to the three points in time event time, reference time, and speech time. Different configurations of these three points along the time line allow a description of different tenses, such as the traditional past perfect, simple past, present perfect, present, simple future, and future perfect.
David Dowty studies the effect of aspectual class on the temporal order of discourse in Chapter 56, ''The effects of aspectual class on the temporal structure of discourse: semantics or pragmatics?'' (pp. 221-244; Source: Linguistics and Philosophy 9, 1986, 37-61). In doing so, his approach is based on (i) the semantic analysis of aspectual class using interval semantics, (ii) a single principle for the interpretation of successive sentences in a discourse (with the principle itself not making reference to the aspectual classes of the involved sentences), and (iii) ''a large dose of Gricean conversational implicature and 'common sense' reasoning based on the hearer's knowledge of the real world information'' (p. 225). In particular, he concludes that the proper explanation of discourse ordering relies to a considerable degree on pragmatics.
Mürvet Enç's aim in Chapter 57, ''Anchoring conditions for tense'' (pp. 245-271; Source: Linguistic Inquiry 18, 1987, 633-657), is to construct a theory of tense that accounts for the interpretation of tenses in embedded clauses. Enç argues against the classical analysis of tense as sentence operator manipulating times in the metalanguage. Her theory proposes that tenses are referential expressions, which she claims stipulates the minimum possible in interpretive rules and accounts for a number of properties of tenses through syntactic conditions governing their interpretation - she purports that the semantic interpretation of temporal expressions is subject to significant syntactic constraints.
Dorit Abusch presents a theory of tense interpretations in main and embedded clauses in Chapter 58, ''Sequence of tense, intensionality and scope'' (pp. 272-289; Source: Proceedings of the West Coast Conference on Formal Linguistics (WCCFL) 7, 1988, 1-14). She claims that the temporal interpretation of sentences is predictable from (i) their logical form (a constraint on which incorporates the notion of transposing context she develops) and (ii) the semantics that is given to tenses. To this end, she discusses the contrast between the sequence-of-tense theory (which assumes the existence of a morphological 'sequence of tense rule' that is responsible for the shift of present tense morphology into past tense in complements of matrix clauses that have past tense morphology) and the independent theory of tense (the embedded tense is always evaluated with respect to the utterance time, independently of the evaluation time introduced by the higher verb).
In Chapter 59, ''Temporal ontology in natural language'' (pp. 290-305; Source: Proceedings of the 25th Annual Meeting of the Association for Computational Linguistics (ACL), Stanford University 1987, 1-7), Marc Moens and Mark Steedman propose a temporal ontology based on the notions of causation and consequence rather than on temporal primitives. They claim that any manageable logic or other formal system for temporal descriptions in natural language will have to comprise such an ontology. The system they devise can explain some of the apparent anomalies and ambiguities that cause problems to approaches to natural-language tenses based on linear models of time.
Dorit Abusch sketches several problems of earlier treatments of tense in embedded contexts in Chapter 60, ''The present under past as de re interpretation'' (pp. 306-318; Source: Proceedings of the West Coast Conference on Formal Linguistics (WCCFL) 10, 1991, 1-12). She proposes that temporal constituents can be captured by a theory of de re interpretation, formalising this idea along the lines of the de re belief theory (David Lewis 1979).
Toshiyuki Ogihara discusses the interaction between tenses, temporal adverbial clauses, and adverbs of quantification such as 'always' and 'often' in Chapter 61, ''Adverbs of quantification and sequence-of- tense phenomena'' (pp. 319-337; Source: Proceedings of Semantics and Linguistic Theory (SALT) 4; Ithaca, NY: CLC Publications, Cornell University 1994, 251-267). He extends a sequence-of-tense rule (which he defines as ''an optional syntactic rule that serves to delete tenses under identity with locally c-commanding tenses'' (p. 319)) to these environments and analyses the scopal relations that emerge.
Chapter 62, ''Quantification over time'' (pp. 338-364; Source: Jaap van der Does and Jan van Eijck (eds.), Quantifiers, Logic, and Language; Stanford, CA: CSLI Publications 1996, 311-336), is the first of four contributions on modality. Henriëtte de Swart extends generalised quantifier theory to cover temporal quantification and interprets adverbs of quantification as generalised quantifiers. Four questions are central to the paper (cf. p. 338): (i) What is the object of quantification of an adverb of quantification? (ii) How do we find the arguments of the adverbial quantifier? (iii) Do the general properties of extension, conservativity, and quantity carry over from the nominal domain? and (iv) Can we establish subclasses of adverbs on the basis of particular semantic properties - and if so, what is their linguistic impact?
Angelika Kratzer characterises the German modal system in Chapter 63, ''The notional category of modality'' (pp. 365-403; Source: Hans- Jürgen Eikmeyer and Hannes Rieser (eds.), Words, Worlds, and Context; Berlin: Walter de Gruyter 1981, 38-74). She asks what the logical nature of interpretations of modals is, what their variety is due to, and how (in the German case) this variety is restricted by the German vocabulary. In the course of the paper, Kratzer proposes a possible-world account of modality in which this notion is interpreted with respect to a conversational background. Modal operators are evaluated with respect to a modal base, that is, a set of accessible worlds, and to an ordering on that set or ordering source.
Donka Farkas' excellent contribution in Chapter 64, ''On the semantics of subjunctive complements'' (pp. 404-435; Source: Paul Hirschbueler and Konrad Koerner (eds.), Romance Languages and Modern Linguistic Theory; Amsterdam: John Benjamins 1992, 69-104), discusses the semantic characterisation of subjunctive argument clauses in Romanian and French. She claims that mood distribution is not random and demonstrates where previous approaches fail, and then proposes an alternative account of modal anchoring in Discourse Representation Theory.
The last contribution in Volume IV, Chapter 65 ''Modal discourse referents and the semantics of the mood phrase'' (pp. 436-461; Source: University of Maryland Working Papers in Linguistics 3, 1995, 224-255) by Paul Portner, pursues the idea that all natural language clauses are modal and involve quantification over a set of possible worlds in the modal base. The modal base of a sentence is context dependent and requires implementing a formal system in which modal bases are treated as discourse referents.
VOLUME V - OPERATORS AND SENTENCE TYPES [ix+433pp.]
Volume V comprises 18 papers on the semantics of three different types of linguistic expression or constructions: adjectives and adjectival modification, negation, and questions, thus triggering insights on the semantics of modification, operators, and sentence types. Corresponding to these themes, the volume is divided into three parts: Part A, 'Adjectives, degrees, and comparatives' (4 contributions), Part B, 'Negation and negative polarity items' (5 contributions), and Part C, 'Questions' (9 contributions).
Part A - 'Adjectives, degrees, and comparatives'
In Chapter 66, ''Two theories about adjectives'' (pp. 7-36; Source: Edward Keenan (ed.), Formal Semantics of Natural Language; Cambridge: Cambridge University Press 1975, 123-155), Hans Kamp outlines a semantics for different types of adjectives (predicative, privative, affirmative, extensional). Instead of conceiving of adjectives as predicates, a treatment of them as functions is discussed. The function approach is more successful than the predicate approach, but still does not allow to capture the semantics of comparatives and superlatives. They require a more sophisticated theory based on supervaluations and graded context-dependent models with probability functions.
Max J. Cresswell presents a non-convincing possible-worlds semantics for English comparative constructions in Chapter 67, ''The semantics of degree'' (pp. 37-69; Source: Barbara Partee (ed.), Montague Grammar; New York: Academic Press 1976, 261-292). He outlines the lambda-categorial languages he bases his semantics on, discusses a semantics of comparative adjectives, mass nouns and plurals, and looks at the metaphysics that underlie comparisons. Then he tackles the ''relation between the underlying formal language of comparatives and their surface form'' (p. 37) - which completes the semantics and syntax for comparatives - and lists a few problem cases, for which he makes tentative suggestions.
Jean-Yves Lerner and Manfred Pinkal's contribution in Chapter 68, ''Comparatives and nested quantifications'' (pp. 70-87; Source: Paul Dekker and Martin Stokhof (eds.), Proceedings of the Eight Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1991, 329-345), is concerned with the relation between comparatives constructions and quantifier terms. Specifically, they discuss the status of the comparative complement as a degree quantifier, the behaviour of quantifier terms occurring in comparative complements, and attributive comparative constructions (in particular the interaction of the comparative operator and the quantifier term in which it occurs). Their analysis relies on the operation of Functional Composition, and primarily on the special case of Nested Quantification.
Chapter 69, ''Comparison and polar opposition'' (pp. 88-106; Source: Proceedings of Semantics and Linguistic Theory (SALT) 7 [not 5 as listed in the volume]; Ithaca, NY: CLC Publications, Cornell University 1997, 240-257) by Christopher Kennedy, is an original and stimulating contribution dealing with the 'cross-polar' anomaly in comparatives (as evident in examples such as #'Mike is shorter than Carmen is tall'), but also generally motivating and developing a specific approach to the semantics of gradable adjectives and the representation of adjectival polarity. Specifically, Kennedy argues that gradable adjectives denote relations between individuals and extents, or intervals in a scale, rather than relations between individuals and degrees, or points in a scale. He furthermore claims that adjectival polarity should be characterised as a sortal distinction between positive and negative adjectives, showing that this approach supports an explanation of cross-polar anomaly as a type of sortal anomaly.
Part B - 'Negation and negative polarity items'
Chapter 70, Gilles Fauconnier's ''Polarity and the scale principle'' (pp. 109-120; Source: Proceedings of the Chicago Linguistics Society 11, 1975, 188-199), aims to show that polarity is only a subcase of a more general phenomenon that can be described in terms of the extended notion 'semantic and pragmatic polarity'. Evidence for this is provided by the following observations: (i) some phrases are polarised with respect to logical structures (they can only occur in sentences associated with logical structures of a certain type), (ii) some phrases are polarised with respect to logical structures and context (their polarity can therefore change from context to context), and (iii) the characteristic polarity-reversal properties of standard polarity items are shared by all elements polarised with respect to logical structures and context. Based on these generalisations, Fauconnier proposes that the distribution of polarised elements follows from the 'scale principle', which he views as a semantic/pragmatic inferential restriction.
Jack Hoeksema discusses a set of monotonicity phenomena in Chapter 71, ''Monotonicity phenomena in natural language'' (pp. 121- 135; Source: Linguistic Analysis 16, 1986, 25-40). He defines and illustrates the notions of direct and inverse monotonicity and indicates how they play a role in the study of inference patterns. Furthermore, Hoeksema explains how the notion of inverse monotonicity makes it possible to state a generalisation about the distribution of negative polarity items.
Chapter 72, the convincing contribution ''Polarity sensitive 'any' and free choice 'any''' (pp. 136-161; Source: Martin Stokhof and Leen Torenvliet (eds.), Proceedings of the Seventh Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1989, 227-251) by Nirit Kadmon and Fred Landman, proposes a unified analysis of the semantic and pragmatic effects of 'any'. Since 'any' can function in two different ways - receiving an interpretation as a negative polarity item (polarity sensitive 'any') and as 'free choice' element - this unified analysis applies to 'any' on both its uses. Essentially, the difference between polarity sensitive and free choice 'any' is reduced to the difference between non-generic and generic indefinites. Kadmon and Landman further show that free-choice 'any' is a generic indefinite with the properties of widening and strengthening, two properties which explain a host of distributional characteristics.
In Chapter 73, ''Nonveridical contexts'' (pp. 162-184; Source: Linguistic Analysis 25, 1995, 286-312), Frans Zwarts draws attention to the semantic properties of what Montague called 'nonveridical contexts'. He is concerned with the relationship between veridicality and monotonicity and also addresses the question of the distribution of polarity sensitive and free choice 'any'. He argues that both uses of 'any' are restricted to nonveridical contexts, supporting the idea that both uses should be analysed as manifestations of a single lexical item (as done in, for instance, Chapter 72). In addition, Zwarts shows that the distribution of negative polarity items in Modern Greek and Romanian is similarly restricted. Non-veridicality also plays a role in the temporal system of many languages. Finally, he claims that recent changes in the use of the Dutch adverb 'ooit' ('ever') can be explained in terms of extensions from monotone decreasing to nonveridical and later veridical environments.
Chapter 74, ''Configurational expression of negation'' (pp. 185-204; Source: Jaap van der Does and Jan van Eijck (eds.), Quantifiers, Logic, and Language; Stanford, CA: CSLI Publications 1996, 203-223) by William Ladusaw, studies the interpretation of negative concord, a morphological property of some languages in which negation is instantiated on several positions in a clause. Ladusaw proposes that the elements of a negative chain should be considered indefinites and argues that such a view of negative concord - one that draws on the theory of indefinites - provides an underlying unity among negative concord terms and negative polarity items, and that this resolves puzzles about how negative concord can be interpreted compositionally.
Part C - 'Questions'
Chapter 75, Lauri Karttunen's ''Syntax and semantics of questions'' (pp. 207-249; Source: Linguistics and Philosophy 1, 1977, 3-44), is the first paper in Part C dealing with the semantics of questions. Karttunen presents a systematic account of the semantics of indirect questions, assuming ''that any adequate solution for them can, in one way or another, be extended to cover direct questions as well'' (p. 208). He discusses previous proposals by, for instance, Hamblin (amongst others). Whereas Hamblin argued for a question to denote the set of propositions expressed by the possible answers to it, Karttunen suggests to make questions denote the set of propositions expressed by their true answers. He then goes on to propose a Montagovian analysis of questions, and devises different rules to generate questions (proto-questions, alternative questions, yes/no questions, etc.).
In Chapter 76, ''Questions, quantifiers and crossing'' (pp. 250-287; Source: The Linguistic Review 1, 1981, 41-79), James Higginbotham and Robert May juxtapose two problems in the syntax and semantics of Logical Form which to them have similar solutions. The first problem concerns an explanation of the source of presuppositions associated with multiple uses of the singular in wh-questions, the second the interpretation of sentences with crossing coreference (sentences with two or more noun phrases, each of which contains a pronoun that may be anaphoric to the other). They claim that these problems can be solved by introducing binary variable-binding operators (binary quantifiers, or binary wh-operators). They give a semantics for both kinds of operators, and propose an operation of absorption that derives n-ary quantifiers from unary ones.
Jeroen Groenendijk and Martin Stokhof explore a possible-worlds account of the question-answer relation in Chapter 77, ''On the semantics of questions and the pragmatics of answers'' (pp. 288-313; Source: Fred Landman and Frank Veltman (eds.), Varieties of Formal Semantics; Dordrecht: Foris 1984, 143-170). They share the view that an adequate theory of questions should deal with direct and indirect questions in a uniform way. A question is treated as a partition of the set of indices, with each element of that partition representing a proposition, a possible semantic answer to the question. In addition, total and partial answers, and direct and indirect questions are treated in this model.
In Chapter 78, ''Towards the semantics of open sentences: wh- phrases and indefinites'' (pp. 314-339; Source: Martin Stokhof and Leen Torenvliet (eds.), Proceedings of the Seventh Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1989, 53-77) by Stephen Berman, the Lewis-Kamp-Heim analysis of indefinites as restricted free variables is extended to the analysis of wh-phrases. Berman first examines parallels in the quantificational behaviour of wh- phrases and indefinites and shows that there are environments where neither type of phrase can be quantified. He argues that a determining role in their quantifiability is played by presupposition accommodation. Secondly, Berman examines asymmetries in the quantifiability of these phrases. He concludes that wh-phrases should also be treated as restricted free variables
Utpal Lahiri studies embedded questions of two types - the 'wonder'- type and the 'know'-type - in Chapter 79, ''Questions, answers and selection'' (pp. 340-352; Source: Proceedings of the North East Linguistic Society (NELS) 21, 1991, 233-246). He claims that the distinction between the predicates of the 'know'-class and the 'wonder'-class cannot be captured by a syntactic distinction. Accordingly, the syntax does not mark the question/answer distinction in the interpretation of embedded interrogatives. Also, the distinction is not a case of selection of different semantic types, but rather seen as a matter of the lexical semantics of the predicates in question. To support this view, Lahiri discusses further evidence from Spanish.
Jonathan Ginzburg presents an account of interrogatives in situation semantics in Chapter 80, ''A quasi-naive semantics for interrogatives and its implications'' (pp. 353-372; Source: Paul Dekker and Martin Stokhof (eds.), Proceedings of the Eight Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1991, 197-212). Questions are treated as unresolved states-of-affairs. Outlining problems for a quantificational view of wh-phrase meaning, Ginzburg also argues for a non-quantificational approach to wh-phrase meaning.
Chapter 81, Veneeta Dayal's ''Two types of universal terms in questions'' (pp. 373-385; Source: Proceedings of the North East Linguistic Society (NELS) 22, 1992, 443-457), discusses questions involving a wh-expression and a universal term. Such questions can be answered in three different ways, by so-called individual answers, pair-list answers, and functional answers. Dayal distinguishes between two types of universal terms and the effects associated with them: the first type of universal terms includes quantificational noun phrases with determiners such as 'each', 'every', and 'both', and the second type of universal terms includes plural definites with determiners like 'the', 'the+numeral', 'those/these', and conjoined proper nouns like 'John and Bill'. She claims that the latter type should be analysed in terms of their individual answers, and that list answers to questions with plural definites represent one of two readings of an individual answer.
In Chapter 82, ''Interrogatives'' (pp. 386-417; Source: Ken Hale and Samuel J. Keyser (eds.), The View from Building 20; Cambridge, MA: MIT Press 1993, 195-227), James Higginbotham advances some parts of a view of interrogative sentences and their interpretations to honour Sylvain Bromberger. One of his major theses is that the semantics of interrogative sentences is given by associating with them semantic objects that he calls 'abstract questions'. To utter the interrogative is then to ask the abstract question it expresses. Elementary abstract questions are seen as partitions of the possible states of nature into families of mutually exclusive alternatives, and complex abstract questions are constructed out of elementary ones. Higginbotham discusses several empirical and theoretical issues related to direct and indirect questions and the licensing of negative polarity items in interrogatives.
The last chapter in Volume V, Chapter 83 ''Interrogatives and polyadic quantification'' (pp. 418-433; Source: Nelia Scott (ed.), Proceedings of the International Conference on Questions; Liverpool: University of Liverpool 1999, 1-14), is written by the editor of the collection, Javier Gutiérrez-Rexach. It develops a treatment of wh-phrases and quantification into questions within the generalised quantifier framework: wh-phrases are uniformly treated as denoting interrogative generalised quantifiers. Quantification into questions is viewed as interaction between a declarative and an interrogative quantifier. This interaction is characterised by polyadic lifts of generalised quantifiers, leading to a variety of polyadic quantifiers.
VOLUME VI - DISCOURSE AND DYNAMICS [ix+483pp.]
The last volume in this collection, Volume VI, deals with discourse structure and information flow, and explores the boundaries between semantics and pragmatics. It contains 18 papers, again divided into three parts: Part A, 'Topic and focus' (4 contributions), Part B, 'Pronouns and anaphora' (8 contributions), and Part C, 'The semantics/pragmatics interface' (6 contributions).
Part A - 'Topic and focus'
In excellently written Chapter 84, ''Topic, focus and quantification'' (pp. 7-32; Source: Proceedings of Semantics and Linguistic Theory (SALT) 1; Ithaca, NY: CLC Publications, Cornell University 1991, 159-187), Barbara Partee explores the possibility of combining aspects of the Prague School of linguistics on topic and focus and other contemporary work on focus-sensitive constructions with the work of Heim and Kamp. In particular, she suggests that topic corresponds to the restrictive term (or domain restriction) and focus to nuclear scope in tripartite structures. She surveys a range of focus-sensitive constructions and concludes that most of them are quantificational in some sense and require something like a tripartite structure for their interpretation. She also analyses several theoretical and empirical consequences and discusses some apparent problems.
The subject of Chapter 85, Manfred Krifka's ''A compositional semantics for multiple focus constructions'' (pp. 33-72; Source: Proceedings of Semantics and Linguistic Theory (SALT) 1; Ithaca, NY: CLC Publications, Cornell University 1991, 127-158), is the development of a framework which allows the formulation of the influence of focus on the semantic and pragmatic interpretation. As basic framework, Krifka deploys the 'structured meaning' approach. (A structured meaning is a pair consisting of a background part and a focus part.) One of the original shortcomings of the apporach have to do with cases involving multiple foci. Therefore, Krifka develops a compositional semantics for multiple focus constructions. He also discusses some extensions and possible problems (among those a combined semantic treatment of focus and topic).
Sjaak de Mey analyses the semantics of focus in generalised quantifier theory in Chapter 86, ''Generalized quantifier theory and the semantics of focus'' (pp. 73-83; Source: Jaap van der Does and Jan van Eijck (eds.), Quantifiers, Logic, and Language; Stanford, CA: CSLI Publications 1996, 269-279). He claims that a semantics for bare focus constructions with 'only' cannot be provided by 'p-sets' (p-sets are determined by all logically possible alternatives) nor by 'salient sets' (sets of contextually determined 'salient' entities). According to de Mey, the most plausible semantics for 'only' is given by generalised quantifier theory, in particular by Boolean Algebra. 'Only' is seen as the superset relation, and as such related to the semantics of other determiners such as 'all'.
In his excellent contribution in Chapter 87, ''Topic'' (pp. 84-109; Source: Peter Bosch and Rob van der Sandt (eds.), Focus, Linguistic, Cognitive, and Computational Perspectives; Cambridge: Cambridge University Press 1999, 142-165), Daniel Büring shows how the formal treatment as developed in Rooth (1985) can be extended to capture not only focus but also sentence topics. He also discusses how the interpretation of focus and sentence topics interacts with the semantics of adnominal quantifiers, yielding a variety of readings (partitive, proportional, and focus-affected).
Part B - 'Pronouns and anaphora'
In the book extract in Chapter 88, ''Pronominal reference: relative pronouns'' (pp. 113-128; Source: Reference and Generality: An Examination of Some Medieval and Modern Theories; Ithaca, NY: Cornell University Press 1962, 108-132), Peter Geach examines certain theories that ascribe reference to pronouns. He observes that the interpretation of pronouns uniformly as variables, following Frege's view, is problematic when considering certain examples such as so- called 'donkey sentences'.
Gareth Evans distinguishes four classes of pronouns in Chapter 89, ''Pronouns'' (pp. 129-157; Source: Linguistic Inquiry 11, 1980, 337- 362). These are (i) pronouns used to make a reference to an object (or objects) present in the shared perceptual environment, or rendered salient in some way; (ii) pronouns intended to be understood as being coreferential with a referring expression occurring elsewhere in the sentence; (iii) pronouns which have quantifier expressions as antecedents and are used as being analogous to the bound variables; and (iv) pronouns which Evans labels as 'E-type pronouns'. He emphasises the differences in interpretation in the fourth class compared with the third class: ''E-type pronouns also have quantifier expressions as antecedents, but they are not bound by those quantifiers'' (p. 130). Evans discusses the existence of E-type pronouns and suggests a semantic description for them. He also proposes a semantics of bound pronouns (pronouns of the third category), and discusses these pronouns in light of the pragmatic theory of co-reference.
Hans Kamp's ground-breaking contribution in Chapter 90, ''A theory of truth and semantic representation'' (pp. 158-196; Source: Jeroen Groenendijk, Theo Jansen and Martin Stokhof (eds.), Truth, Interpretation and Information; Dordrecht: Foris 1984, 1-41), introduces a new theory of sentence and discourse interpretation that addresses the main problems in the interpretation of pronouns and indefinites in donkey sentences. This theory, called Discourse Representation Theory (DRT), is an attempt to combine a definition of truth with a systematic account of semantic representation (involving a constructional, non-compositional approach to meaning) and is thus essentially a dynamic theory of meaning. Interpretations are mediated by the construction of Discourse Representation Structures, which in turn are embedded in models. Kamp shows that the theory provides: (i) a general account of the conditional; (ii) a general account of the meaning of indefinite descriptions; and (iii) a general account of pronominal anaphora.
In Chapter 91, ''Modal subordination and pronominal anaphora in discourse'' (pp. 197-233; Source: Linguistics and Philosophy 12, 1989, 683-721), Craige Roberts studies what she refers to as 'modal subordination' and how it affects and restricts anaphoric relationships. First, she relates the notion of the mood of a sentence to a theory of modality in model theoretic semantics, showing how this is relevant for modal subordination. Then, she discusses the relevance of anaphora for a theory of modal subordination, and proposes an enrichment of Discourse Representation Theory to deal with modal subordination and generalised subordination in discourse, including cases of what she calls 'telescoping'.
Chapter 92, Paul Dekker's ''Existential disclosure'' (pp. 234-259; Source: Linguistics and Philosophy 16, 1993, 561-587), presents a system of dynamic interpretation. This system is based on Groenendijk and Stokhof's (1990) Dynamic Montague Grammar, in which the Montague grammar is adapted in order to incorporate Discourse Representation Results. Dekker suggests that dynamic interpretation is connected with the possibility of an operation he calls 'existential disclosure' - the possibility of addressing (dynamic) existentially closed (implicit) arguments as if they were free variables. He aims at showing how existing treatments of relational nouns, adverbial modification and tense in discourse can be expressed within such a dynamic framework.
Gennaro Chierchia's book extract in Chapter 93, ''Dynamic binding'' (pp. 260-282; Source: Dynamics of Meaning; Chicago, Ill.: University of Chicago Press 1995, 62-84), discusses two different readings of donkey sentences: the existential reading and the universal reading, the existence of which is not without problems for classical Discourse Representation Theory. Chierchia proposes first an extension of Discourse Representation Theory in which adverbs of quantification are not unselective and can choose the arguments they quantify over. He integrates his results in a dynamic version of Montague's Intensional Logic which he calls the 'Dynamic Binding' approach. (Note that the book chapter is not reprinted in full.)
Jeroen Groenendijk, Martin Stokhof and Frank Veltman's aim in Chapter 94, ''Coreference and contextually restricted quantification'' (pp. 283-302; Source: Proceedings of Semantics and Linguistic Theory (SALT) 5; Ithaca, NY: CLC Publications, Cornell University 1995, 112-129), is to argue that update semantics is a natural framework for contextually restricted quantification. (In update semantics the meaning of a sentence is identified with its context change potential, with contexts being taken as information states.) They furthermore illustrate its use in the analysis of anaphoric definite descriptions and suggest to treat anaphoric definite descriptions - together with certain other anaphoric terms - as quantifiers (where quantification is dynamic and contextually restricted).
Chris Barker tackles the 'proportion problem' in Chapter 95, ''A presuppositional account of proportional ambiguity'' (pp. 303-319; Source: Proceedings of Semantics and Linguistic Theory (SALT) 3; Ithaca, NY: CLC Publications, Cornell University 1993, 1-18). The 'proportion problem' is a systematic ambiguity associated with proportional adverbial quantification such as 'usually' and 'mostly'. Since one of the readings often seems to be strongly preferred, Barker explores the factors that constrain which of several construals of a proportional sentence will be available for a token uttered in a particular context. He proposes that there is a conventional implicature connected to the use of proportional adverbial quantifiers, and that different proportional readings give rise to distinct presuppositions.
Part C - 'The semantics/pragmatics interface'
Robert Stalnaker proposes how content and context might be represented in a theory of speech in Chapter 96, ''Assertion'' (pp. 323- 340; Source: Peter Cole (ed.), Syntax and Semantics 9; New York: Academic Press 1978, 315-322). He makes some claims about the way assertions act on the contexts in which they are made, and about the way contexts constrain the interpretation of assertions. Three principles are mentioned, which illustrate the interaction of context and content and which can be defended as essential conditions of rational communication. These principles - discussed by Stalnaker in turn - are: (i) a proposition asserted is always true in some but not all of the possible worlds in the context set; (ii) any assertive utterance should express a proposition, relative to each possible world in the context set (and that proposition should have a truth value in each possible world in the context set); and (iii) the same proposition is expressed relative to each possible world in the context set.
Chapter 97, David Lewis' ''Scorekeeping in language game'' (pp. 341- 358; Source: Rainer BŠuerle, Urs Egli and Arnim von Stechow (eds.), Semantics from Different Points of View; Berlin: Springer 1979, 172- 187), proposes to model conversational interaction in terms of a game metaphor. Presuppositions evolve according to rules of accommodation. Such rules specify, for example, that any presuppositions that are required by what is said come into existence provided that nobody objects.
Chapter 98, Enric Vallduvi's ''A theory of informatics'' (pp. 359-384; Source: The Informational Component; PhD diss., University of Pennsylvania 1990; Ann Arbor, MI: Garland 1992, 66-94), is an extract from the author's PhD dissertation. It analyses information packaging as part of an autonomous module that Vallduvi calls 'informatics'. The chapter provides an account of the role of informatics in the larger language apparatus. Speakers have a knowledge store and elements are 'filed' into that store according to their informational role focus or ground. Vallduvi discusses several phenomena from English and Catalan, showing how the linguistic encoding of information packaging can be different in each language.
Chapter 99, ''The context-dependency of quantifiers'' (pp. 385-397; Source: Restrictions on Quantifier Domains; PhD diss., University of Massachusetts at Amherst 1995, 27-36), supplies another extract from a PhD dissertation - namely Section 2.2 (and thus not even a whole chapter). Due to this extract being taken out of a bigger whole, the extract's aim and its relevance do not become clear. Kai von Fintel studies the context dependence of quantifiers and claims that quantifiers are restricted by resource-domain arguments.
Chapter 100, Dov Gabbay and Ruth Kempson's ''Natural-language content: a proof-theoretic perspective. A preliminary report'' (pp. 398- 428; Source: Paul Dekker and Martin Stokhof (eds.), Proceedings of the Eight Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1991, 173-195), reports on work in progress, work that is aimed at building a formal model of the process of utterance interpretation from a procedural perspective. Gabbay and Kempson introduce the background logical framework, sketch out the account itself, apply it to anaphora and wh-binding, and show how it affects underlying assumptions about natural language.
The last chapter not only in this volume, but in the whole collection - Chapter 101, ''Mathematical treatment of discourse contexts'' (pp. 429- 454; Source: Paul Dekker and Martin Stokhof (eds.), Proceedings of the Tenth Amsterdam Colloquium; Amsterdam: ILLC, University of Amsterdam 1995, 21-40) - presents a theory of discourse structure and hence interpretation which integrates semantic and pragmatic information. Nicholas Asher reviews and lays out the Segmented Discourse Representation Theory (SDRT), which like DRT itself exploits representations to model discourse contexts and to determine the semantic effects of discourse structure. Asher first gives some applications of SDRT to problems of discourse interpretation. He then develops an alternative model-theoretic approach to discourse contexts for a theory of pragmatics and semantics, and finally establishes an equivalence of this approach to SDRT.
The collection concludes with the Index (pp. 455-483), which merges both name and subject index. Unfortunately, the index is not listed throughout the collection (not in any table of contents), so it came rather as a surprise after I had read all the papers and could surely have used the cross-referencing potential of an index in the course of reading.
It cannot be the aim of this evaluation to discuss single contributions. Therefore, I will merely comment on the overall collection as such.
The collection is a commendable attempt to bring together influential papers from formal semantics, many of which have only been published in conference proceedings of narrow diffusion. It is a specialised anthology targetted at semanticists who are have some acquaintance with the tools and methods of formal semantics. In particular, it does not contain and cannot fulfill the function of an introduction to the concepts of formal semantics, and will thus not be easily accessible to semanticists from other theoretical backgrounds, let alone to scholars in other areas of linguistics or students.
The editor, as stated in the preface, selected the contributions specifically from the field of formal semantics. On this basis, the collection's title is misleading. It should have been ''Formal Semantics. Critical Concepts in Linguistics'', in order to not even indicate a potential coverage of the field of semantics. An overall coverage would have to include other theoretical approaches and work by other important semanticists, such as Jackendoff, Cruse, Fillmore, Talmy, Pustejovsky, Wierzbicka, and many more. Also, important semantic subfields such as lexical semantics are generally not focused on in formal semantic and are hence underrepresented in this anthology.
A major drawback of the volumes is the high number of misprints, which are specifically distorting in formula and decrease the readability immensely. Often one feels the need to consult the original for comparison. Given the nature of the misprints, I am tempted to speculate that the character recognition results following the scanning of a camera-ready copy were not corrected any more. Together with instances as the one outlined for Chapter 46, this is disappointing, also given the high price of the volumes.
As a collection of formal semantic papers the present anthology certainly brings together most of the highly influential papers, written by leading scholars in the field, and makes papers readily available that were not so easily accessible before. An inclusion of work by, for instance, Donald Davidson, Alfred Tarski, or Mats Rooth would have made the collection even more complete, but I am aware that not every important contribution to formal semantics can be included in such a collection.
REFERENCES (for both parts)
Benthem, Johan van 1986. Essays in Logical Semantics. Dordrecht: Reidel.
Groenendijk, Jeroen and Stokhof, Martin 1990. 'Dynamic Montague Grammar.' In: László Kálmán and László Polós (eds.), Papers from the Second Symposium on Logic and Language, Akadémiai Kiadó, Budapest, 3-48.
Jespersen, Otto 1924. The Philosophy of Grammar. London: George Allen and Unwin.
Kamp, Hans and Reyle, Uwe 1993. From Discourse to Logic. Dordrecht: Kluwer.
Katz, Jerrold J. and Fodor, Jerry A. 1963. 'The structure of a semantic theory.' In: Language 39, 170-210.
Lewis, David 1979. 'Attitudes de dicto and de re.' In: Philosophical Review 88: 513-543.
Liu, Fengh-Hsi 1990. Scope Dependency in English and Chinese. PhD diss., UCLA.
May, Robert 1977. The Grammar of Quantification. PhD diss., Cambridge, MA: MIT.
Rooth, Mats 1985. Association with Focus. PhD diss., University of Massachusetts, Amherst.
Zaefferer, Dietmar 2002. 'Polysemy, polyvalence, and linking mismatches. The concept of RAIN and its codings in English, German, Italian, and Spanish.' In: DELTA - Documentação de Estudos em Lingüística Téorica e Aplicada 18 (spe.), 27-56. Special Issue: Polysemy.
Andrea C. Schalley is Australian Research Council Postdoctoral Fellow at the University of New England, Australia. Trained in theoretical linguistics (PhD University of Munich, Germany, 2003), she is primarily interested in semantic representation frameworks, questions of modelling adequacy in theories about language, the relationship between language and cognition, and constraints on linguistic variability.