The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.
The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin
Woods, Michael (2003) Conditionals, Oxford University Press.
Announced at http://linguistlist.org/issues/14/14-1764.html
Maria Leonor Santos, Universidade Federal da Paraíba/ Universidade Federal de Santa Catarina.
Synopsis of the book
Michael Woods' Conditionals should have been part of a more comprehensive book on Philosophical Logic, that was left incomplete. The author died in April 1993, and his Brasenose colleague John Foster is responsible for the transcription, ordering and editing of the material left. The object of this review is the first paperback edition of Conditionals (2003).
The present book contains the Editor's Preface, by David Wiggins, 8 Chapters on conditionals, by Michael Woods, a Commentary by Dorothy Edgington, a List of Works Cited, an Obituary by John Ackrill, followed by Michael Woods Curriculum Vitae and the Index.
Conditionals are, as Editor David Wiggins says in his Preface to this book, ''one of the oldest, most troublesome questions in logic''. As it is widely known, it is quite common for us to find problems when trying to accommodate the description of natural language connectives to their (presumed) corresponding formal connectives. Natural language connectives do not usually behave exactly the way the formal connectives do, as it is easy to see when we contrast uses of ''and'', ''or'', and ''not'', on the one hand, with predicate calculus formulae containing formal conjunction, disjunction and negation, on the other. Conditionals, however, beat them all, and present a sort of acute case of discrepancy.
Chapter One (The Varieties of Conditionals) presents the problem and introduces the two types of conditionals, as they are usually known: the so called indicative conditionals, on the one hand, and the subjunctive, or counterfactual conditionals, on the other. The first problem with conditionals is that they seem to be a form of sentence composition, but it is very difficult to determine whether the form is a truth-functional one, and if so, what function gives us the truth-value of the complete sentence, from the truth-values of the component sentences. Besides, conditionals come in assertions, but also in questions and commands (a fact that is not always taken into consideration in other treatments of the subject). Non- assertoric conditionals are postponed to Chapter Seven. Woods also discusses the inadequacy of the terms ''subjunctive'' and ''counterfactual'' for conditionals, and the widespread characterization of the counterfactual as one that implies the falsity of the antecedent. He surveys briefly some of the solutions found in the literature to the question of assertibility of conditionals, and introduces his own taxonomy: those conditionals that are made up of two stand-alone sentences plus the ''If... then...'' connective are the Simple Conditionals, and those that cannot be thus analyzed are called ''Counterfactual'' Conditionals, for lack of a better term.
Chapter Two (Theories of Simple Conditionals) develops an exposition of several proposals already mentioned in Chapter one. Briefly, Simple Conditionals may be regarded as always having a truth-value, or as not (or not always) having a truth-value. If Simple Conditionals are regarded as truth-functional, then they are not always assertible, even in cases when they are true. That is, if we take the logic conditional (also called the material condition, or material implication) as the description of the truth- values of the Simple Conditional, then we need another parameter, that of assertibility, to account for the divergence between truth and use. At least two problems arise in connection with the notion of assertibility: it seems to be especially difficult to state for conditionals, and it is also difficult to distinguish from the acceptability of the utterance containing conditionals. Five theories for handling the matter are then presented and discussed in a preliminary way. Very briefly, they are:
i) Simple Conditionals (in natural language) have truth- values in the same the way material conditions (in formal logic) do, and the difference between the truth-values of Simple Conditionals and their assertibility is explained in terms of pragmatics.
(ii) Simple Conditionals have truth-values in the same the way material conditions do, and the difference between truth-value and assertibility is a difference in meaning.
(iii) Simple Conditionals have truth-values, but not always, and do not follow the truth-table for material conditions, but rather:
- when P and Q are both true, ''If P then Q'' is true
- when they are both false, ''If P then Q'' is false
- otherwise, there is no truth-value.
This is a consequence of regarding ''If P then Q'' as a conditional assertion, that is, an assertion of Q on the condition that P. If P is not the case, there is no assertion, and no truth-value.
(iv) Simple Conditionals have truth values but are not truth-functional. The truth of the material condition may be necessary but not sufficient for the truth of the conditional. ''If P then Q'' is regarded as a metalinguistic assertion.
Problems and inadequacies of each theory are also discussed.
Chapter Three (Ramsey's Test and Adam's Hypothesis) is dedicated, as the titles states, to discussing two approaches already mentioned. Ramsey and later Adams discussed the form ''If P then Q'' as an assertion of Q on the condition that P. Thus, someone accepting that Q on the condition that P is described as making minimal revisions in his or her stock of beliefs, and conditionals are then regarded as epistemic. The notion of minimal revision, and the calculus of conditional probability are then discussed. According to Woods, Adams's Hypothesis seems to give a better account (better than, for example, the use of the notion of relevance) of the behavior of conditionals in certain troublesome cases. The troublesome cases would be the failure of Simple Conditions to comply with transitivity, contraposition, strengthening of the antecedent, the equivalence of ''P or Q'' with ''not If P then Q'', and with modus tollens.
Chapter Four (Simple Conditionals and Truth-Values: Some Proposals) brings a review of what has been studied so far, and narrows down the available options for treating Simple Conditionals. Woods discusses Adams's Hypothesis still further, and also Lewis's epistemic approach, Jackson's notion of robustness, and Grice's pragmatic approach.
Chapter Five (Conditionals and Possible Worlds). In Chapter one, Woods denied the usual characterization of ''Counterfactual'' Conditionals as those that imply the falsity of the antecedent. In this chapter, he applies a widespread approach to counterfactuality (possible worlds), due mainly to the work of Lewis and Stalnaker, to the analysis of Simple Conditionals. Woods discusses the apparatus of possible worlds, the problem of ordering and selecting worlds, and the solutions given by Lewis and Stalnaker. The application of possible worlds to Simple Conditions gives a good explanation to a wide variety of cases, but brings several other problems into the field, like the adequacy of the notion of similarity between worlds, and whether the choice of worlds should be guided by causal laws (as proposed by Bennett) and not by similarity. Even if we accept an approach based on possible worlds, it cannot give a proper explanation to the reasons we have for using conditionals the way we do. It does not seem possible to explain why the form of Simple Conditional should be attached to the meaning it has, and it seems to be misguided to apply a possible world treatment to Simple Conditionals.
Chapter Six (Compound Conditionals and Truth-Values) examines four types of case in which a conditional is embedded in a larger sentence. Woods presents his arguments for preferring to regard conditionals as conditional assertions, lacking truth-values. The four cases are:
(i) a conditional consequent
(ii) a conditional with disjunctive antecedents
(iii) a disjunction of conditionals
(iv) a conditional with a conditional antecedent, and the negation of a conditional, which present the greatest problems.
Chapter Seven (Theory of Simple Conditionals; Non- Assertoric Conditionals) states the two remaining approaches for conditionals:
(i) Simple Conditionals do not have truth-values. They are used to make conditional assertions
(ii) Simple Conditionals do have truth-values, but they are relative to the epistemic state of the speaker.
The second view, however, may be regarded as leading to greater complexity and not offering any advantages in turn, and the author prefers to consider Simple Conditionals as lacking truth-values. They are used to make conditional assertions, and one extra advantage of this point of view is the possibility of describing non-assertoric conditionals (questions and commands) as parallel cases.
Chapter Eight (Sketch of a Theory of ''Counterfactual'' Conditionals) offers an account of the ''Counterfactual'' Conditionals in terms of epistemic states of the speaker, as it was already presented in chapters three and four, concerning Simple Conditionals. This approach should allow us to account for several types of sentences that are not usually grouped with the ''Counterfactual'' type. It accommodates the problems of temporal asymmetry, ordering of worlds, choosing of worlds, and also the detailed interplay between the tense of verb used and the interpretation of the sentence. The concluding remarks stress once more the fact that even if possible worlds are adequate to describe the form of this kind of conditionals, they cannot help us to explain the use we make of conditionals, and give too much attention to ''fantasies'' devoid of more immediate practical reasons.
The Commentary by Dorothy Edgington takes up roughly one third of the book. It is a detailed response to several of the topics of the book, providing, as Dorothy Edgington says, background for some of the ideas presented, extra arguments in favor of some positions, and even some disagreement (concerning the treatment of ''will'' sentences).
An obvious drawback of this book is that it was left incomplete, and the text is rather packed, with a lot of theory pressed together. At least one section is incomplete: Chapter seven, on compound conditionals. Also, the titles of the chapters do not strictly correspond to the subjects presented in them: the discussion goes far beyond what is mentioned in the title, and that is a little misleading. And the reader, however sympathetic to the cause, could probably use some more examples, for the subject is indeed a hard one. Then there are the good aspects: for a book that was left incomplete, this is an exceptionally clear and well-written one, especially one on such a demanding subject. Each new chapter (apart from the first) begins with a summary of the ideas discussed in the preceding one, and ends with a sketch of what will be presented in the next, thus making it a very readable essay, and not at all a collection of loose theories. For the beginner in the field, the footnotes provide a good introduction to the literature on the subject. Also on the plus side are the Editor's notes and the Commentary. In this rather small book, then, Michael Woods managed to present us with a good map of the area: a definition of conditionals, a sketch of the reasons why they are so difficult to deal with, a survey of the main theories proposed to sort them out, and some fresh proposals of his own, which is reason enough for us to read and enjoy his book.
ABOUT THE REVIEWER:
Maria Leonor Santos teaches Linguistics at Federal University of Paraíba, Brazil, and is now working for her thesis on conditionals (in Brazilian Portuguese) at Federal University of Santa Catarina. Her main interests are Logic, Lexical Semantics, and History of Linguistics.