This book "supplies a vocabulary of English words and idiomatic phrases 'arranged … according to the ideas which they express'. The thesaurus, continually expanded and updated, has always remained in print, but this reissued first edition shows the impressive breadth of Roget's own knowledge and interests."
Katja Jasinskaja, Centre for General Linguistics (ZAS), Berlin
This book presents a comprehensive discussion of quantity implicatures, which are a family of pragmatic inferences that, in the pioneering work of Paul Grice, were analysed as resulting from the assumption that participants in a conversation observe the Maxim of Quantity, i.e. provide as much (relevant) information as they can (without lying, making unwarranted statements, etc.) The book covers all the standard varieties of quantity implicature, including scalar implicature, exhaustivity, and conditional perfection. It also covers free choice inferences, which the author shows to follow the same pattern of reasoning. Since Grice's original proposal, the same class of phenomena received a large number of alternative analyses, including lexicalist, syntactic and semantic accounts, as well as pragmatic accounts in other flavours. This book presents a vigorous defence of a radical pragmatic view on the matter and argues essentially for ''a return to Grice'' - in spirit, if not in letter. One part of the argument is based on considerations of theoretical parsimony: if it is possible to account for some part of content conveyed by natural language sentences as an inference based on general assumptions of human rationality and cooperativity, rather than making additional stipulations in the grammar or the lexicon, then one ought to do so. The second part of the argument consists of showing that many phenomena related to quantity implicatures that were believed to resist a Gricean account can in fact be given one, if a number of details in the Gricean picture are set straight. The book culminates in a thorough discussion of one such phenomenon, the so-called embedded implicatures, i.e. cases where it looks as if a quantity-based inference takes place in the scope of some operator in the sentence.
Apart from a short introduction and afterword, the book consists of eight chapters. The first five lay out the foundations of the author's version of Gricean pragmatics, positioning it among previous proposals of both Grice himself and others, and addressing some fundamental issues and criticisms of the framework. Chapter 1 presents the main concepts introduced by Grice, marking the points in which the author endorses or deviates from the orthodox picture. Chapter 2 introduces what the author calls the ''Standard Recipe'': a pattern of reasoning standardly assumed to be the Gricean solution to quantity implicatures, which largely amounts to the exclusion of more informative alternative statements the speaker could have made but didn't. The author emphasises that many misconceptions about what Gricean pragmatics can and cannot do result from overlooking two points: first, that the standard recipe is a form of abductive reasoning (reasoning to the best explanation) and not a logical calculus, and second, that a sharp distinction should be drawn between weak and strong implicatures (i.e. inferences to the effect that the speaker does not believe P vs. that he/she believes not P). The latter follows the derivation of the former on the assumption that the speaker is competent. Chapter 3 focuses on scalar implicatures, which are best known to bear suspicious resemblance to inferences triggered by specific lexical items (e.g. some, or, warm, etc.) and have been the main source of inspiration for lexicalist accounts. Chapter 4 replies to charges of psychological inadequacy frequently put forward against Gricean pragmatics. The author brings experimental evidence from published psycholinguistic, psychological, and brain studies in defence of the Gricean theory as a theory of experimental pragmatics. Finally, Chapter 5 goes back to scalar inferences and summarises existing experimental evidence that opposes the view that scalar inferences are lexically triggered defaults.
The last three chapters of the book go into a detailed discussion of a range of phenomena that the author argues to belong to the class of quantity implicatures. Some of these phenomena have been traditionally seen as problematic for the Gricean theory, most notably free choice inferences and embedded implicatures. The author fleshes out his version of Gricean pragmatics and puts it to the test on these phenomena. Chapter 6 argues for an ''intentionalist turn'' in Gricean reasoning. That is, instead of reasoning directly with alternative statements (as one would do according to the ''standard recipe''), one reasons with hypotheses about alternative possible ''states of mind'' of the speaker, and excludes such hypotheses if the respective state of mind would make the speaker utter something different from what he/she did. One of the advantages of this approach is that it does not generate inferences that are inconsistent with the propositional content of the sentence ('what is said') since the latter functions as a constraint on the speaker's possible state of mind. This problem arises with free choice inferences under the ''standard recipe'' reasoning pattern, but disappears on the intentionalist view, as the author demonstrates. The approach is further applied to varieties of quantity implicature that typically involve an indefinite number of alternative statements - exhaustivity and conditional perfection - for which it seems to account in a computationally more viable way. Chapters 7 and 8 deal with embedded implicatures. Chapter 7 is a crusade against all kinds of non-pragmatic (i.e. lexicalist or syntactic), or as the author labels them, 'conventionalist', accounts of this class of phenomena. One part of the argument concerns cases where conventionalist theories fail to make any predictions: (a) weak implicatures (see above); and (b) inferences at a level higher than a single sentence, or discourse-level implicatures. The latter are unproblematic for the Gricean theory, since it works with speech acts, which can be complex and consist of other speech acts. The second part of the argument concerns false predictions of conventionalist approaches, or cases where embedded implicatures are predicted to appear but don't. The main source of evidence is the much debated experimental study of Geurts and Pouscoulous (2009a), which finds strikingly low numbers of embedded implicatures in intensional and quantificational contexts. Finally, Chapter 8 presents the positive part of the proposal. The problematic facts related to embedded implicature are divided into two groups. For the first group, which includes disjunctions, sentences with indefinites, belief reports and factive predicates, the author develops analyses that derive what looks like an embedded implicature by purely Gricean means without postulating an ''implicature operator'' in the sentence structure. The second group is embedded implicatures in contrastive contexts. The author argues that these cases do involve local pragmatic strengthening of the literal meaning of subsentential linguistic expressions, but these inferences are exceptional in many respects and are not implicatures in the author's terminology, even though reasoning with Gricean principles plays a role in their derivation as well.
The main conclusion of the book is that after some amendments to the Gricean theory of pragmatics, there is no need to relegate the phenomena studied in this book to the domain of linguistic convention, and since considerations of theoretical parsimony put the Gricean approach at an advantage from the start, the odds are in Grice's favour.
This book is excellent. This applies to both presentation and content.
The problems, the positions, and the arguments are brilliantly presented, in a clear, step-by-step fashion that is easy to follow. The book is not long - just above 200 pages - yet the author succeeds in saying a lot within that limited space. The presentation is dense, but not too dense in the sense that one has to reflect on every word in order to understand the content. It is more like every page is worth reading. The chapters are proportionally short, so the information is cut up into very manageable pieces.
The book is written for a broad audience of philosophers, linguists and psychologists with even minimal exposure to logic and formal languages. One encounters an occasional phi and psi, but not a single ''difficult'' formula. At the same time, all the assumptions and inference steps are stated in a very systematic and explicit way, so one gets the impression that formalising the proposed approach should not be that difficult. For this reason, I think the book is very suitable for students as an introduction to Gricean implicatures (that is, not counting the fact that it is not ideologically unbiased). I would definitely recommend Chapter 2 as preparatory reading for a class on conversational implicature. Finally, the text is also sprinkled with sardonic humour, mostly directed at the author's theoretical opponents, which gives the discussion at times a somewhat polemic tone, but makes it a joy to read.
Concerning content, I think the book is a significant contribution to pragmatic theory. In terms of data coverage, it is an improvement on Grice, and in terms of beauty and simplicity (largely due to the recognition of Grice's original ideas), it is an improvement on most theoretical work in linguistic pragmatics of the last decades. The book presents serious discussion of some burning issues in the field, which also means that the author's proposals are far from uncontroversial. The most central theme of the book, embedded implicatures, is a battlefield a civilian shall not enter without a bulletproof vest. Therefore, I will refrain from a critical evaluation of this part of the book and point to the lively discussion in the field's leading journals and elsewhere, featuring Chemla (2009), Geurts & Pouscoulous (2009b), Sauerland (2010), Ippolito (2010), Clifton & Dube (2010), Chemla & Spector (2011), Geurts (2011), and Gajewski and Sharvit (2011) (not to imply that the list is exhaustive).
Instead of commenting on what everyone is commenting on, I will make some remarks on issues that tend to attract less attention. One of the arguments against ''implicature operators'' in the sentence structure put forward in the book (see pp. 23-24 and 154) is the observation that sequences of more than one sentence seem to give rise to implicatures in exactly the same way as single sentences do. I agree with the basic intuition, and it is high time to raise the issue. However, I am not convinced that applying Gricean reasoning to suprasentential units is the only way to capture these intuitions. Just like embedded implicatures in the author's proposal, a large class of discourse-level quantity implicatures might be reducible to the sum of sentence-level implicatures under a number of additional (but independently motivated) assumptions. In example (54) from Chapter 7, the intuition that this discourse, when used as an answer to the question, ''Which places did you see on your trip to Italy?'', implicates that the speaker didn't see other places in Italy except Naples, Rome, and Ravenna could be explained by the following line of reasoning: The speaker divides the question explicitly asked into two implicit subquestions; ''Which places did you see with Julius?'' and ''Which places did you see without Julius?''. The first sentence answers the first subquestion and implicates that the speaker didn't see, let's say, Venice with Julius. Similarly, the second sentence implicates that the speaker didn't see Venice without Julius. But if the speaker neither saw Venice with nor without Julius, then she didn't see Venice. So the discourse-level implicature simply follows from the conjunction of sentence-level implicatures.
(54) Julius and I first went to Naples and Rome together. Then, while he went to see Milan, I visited Ravenna.
A crucial assumption in this argument is that the domains of the subquestions split up the domain of the original question in an exhaustive way. However, a constraint, or at least a preference along these lines follows from common assumptions about what is a good discourse strategy in many question-based approaches to discourse structure (e.g. Roberts 1996). This is, of course, good news for a conventionalist because, at least in cases like (54), it would be enough to make your ''implicature operator'' sensitive to a contextual parameter that imports the domain restriction of the question under discussion and to let the theory of discourse structure do the rest. Importantly, this is not intended as an attempt to dismiss the idea of Gricean reasoning at the discourse level. There are more complex examples than (54), and the requirement of exhaustive split is most probably a violable constraint, so it should be more revealing to look at cases when this requirement is violated. That is to say, the discourse-based argument against conventionalism is not as simple and straightforward as it is presented in the book, and as the author correctly points out, the issue urgently demands proper attention from researchers.
Another comment concerns embedded implicatures in contrastive contexts (see pp. 140-143, 181-189) and the distinction between sentential and subsentential speech acts (e.g. the act of referring, see p. 25). Contrastive contexts are the only case where the author concedes that pragmatic strengthening of the literal meaning of an expression can take place locally (i.e. also in the scope of operators). He also agrees that this pragmatic strengthening is the result of Gricean reasoning. However, he refuses to call that an implicature. My feeling is that the discussion of this issue is cut a bit too short in the book. In the end, one does not get rid of the impression that the question might be purely terminological. Without a more systematic discussion of the “real” differences between pragmatic inferences associated with sentential and subsentential speech acts, the author's narrow understanding of the term ''implicature'' looks somewhat artificial. On a broader understanding, embedded implicatures would not be a contradiction in terms, but an exceptional and rare phenomenon.
To conclude, ''Quantity Implicatures'', by Bart Geurts, is a wonderfully carried out, thought-provoking theoretical study of the advertised quantity implicatures. It is a good read despite its dry subject matter, and a must for anyone interested in theoretical pragmatics.
Chemla (2009). Universal implicatures and free choice effects: Experimental data. Semantics and Pragmatics, 2(2):1--33.
Chemla & Spector (2011). Experimental evidence for embedded scalar implicatures. Journal of Semantics.
Clifton & Dube (2010). Embedded implicatures observed: A comment on Geurts and Pouscoulous (2009). Semantics and Pragmatics, 3(7):1--13.
Gajewski & Sharvit (2009). In defense of the grammatical approach to local implicatures. Natural Language Semantics.
Geurts (2011). Embedded implicatures (cont.). Talk presented at the DGfS Workshop on Implicatures and Discourse Structure.
Geurts & Pouscoulous (2009a). Embedded implicatures?!? Semantics and Pragmatics, 2(4):1--34.
Geurts & Pouscoulous (2009b). Free choice for all: A response to Emmanuel Chemla. Semantics and Pragmatics, 2(5):1--10.
Ippolito (2010). Embedded implicatures? Remarks on the debate between globalist and localist theories. Semantics and Pragmatics, 3(5):1--15.
Roberts (1996). Information structure in discourse: Towards an integrated formal theory of pragmatics. OSU Working Papers in Linguistics, 49:91--136.
Sauerland (2010). Embedded implicatures and experimental constraints: A reply to Geurts & Pouscoulous and Chemla. Semantics and Pragmatics, 3(2):1--13.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Katja Jasinskaja is a researcher at the Centre for General Linguistics
(ZAS), Berlin, working on the project 'Implicatures and Discourse
Structure'. Her main research interests lie in the area of discourse
theory, pragmatics and semantics.