The book by K. M. Jaszczolt under review here is concerned with the representation of time both in language and in linguists' analyses of language. Insofar as the linguistic representation of time is mediated by the human concept or concepts of time, the monograph is necessarily concerned with temporal phenomenology as well. For anyone interested in this range of issues, the volume is extraordinarily rich, for all its brevity. That is not to say that every philosopher or every linguist will find this dense and quite challenging book engaging.

For the most part, linguists' analyses of temporal semantics are formulated in terms of what the author calls ‘real’ time: the ordered sequence of moments along the line of time, McTaggart's (Reference McTaggart1908) famous B-series. Thus the definitions of tenses in linguistics have usually involved the relative ordering of eventualities. But the kind of time that concerns humans most is McTaggart's A-series, the speaking of time in terms of past, present and future or a set of temporal positions that are in constant flow.

For those who conceive of language as representing reality indirectly, mediated by psychological conceptualization, this poses a challenge. Just how does the A-series concept of time, what Jaszczolt calls psychological, conceptual, or ‘internal’ time, relate to the B-series ‘real’ time, and how (and to what extent) does language represent ‘real’, as opposed to ‘internal’ time? After all, the whole point of language is, ultimately, to represent, and thereby to enable its users to function effectively in, the real world.

This does indeed raise significant and central questions in the phenomenology of time, but it is questionable whether these are significant questions for the semantics of tense, or, if they are, just what consequences they have for the workaday scholar of tense and aspect.

Truth-conditional semantics assumes that linguistic semantics is ultimately about truth, interpreted as correspondence to the real world. But a good case could certainly be made that language does not represent the ‘real’ world, and hence semantics cannot be about correspondence with it; rather, language, insofar as it represents time, represents a conceptual time that has little more in common with ‘real’ time than the name. The reality of what Jaszczolt calls ‘real’ time is especially in question when we consider that, as an idealized abstraction from quotidian phenomena, the physicists' ‘time’ may simply be an alternative conceptualization to the naïve ‘internal’ time.

Assuming that both the question of the relationship of ‘real’ time to ‘internal’ time and that of the representation of ‘real’ time in language are equally meaningful and likewise significant for the linguist, the author argues that the answers to both questions involve modality, as the subtitle of her monograph reveals.

That there is some sort of connection between temporality and modality has long been suspected by linguists. There are obvious associations between the markers of the two categories. Sigmatic stems in Indo-European, for example, occur in both the subjunctive and the future (hence Hahn Reference Hahn1953). Modal operators in many languages have temporal implications, especially as regards the future (Palmer Reference Palmer1979). Irreality and counter-factuality are commonly marked by temporal markers such as the past tense (Steele Reference Steele1975, James Reference James1982).

Consequently, given the dual temporal/modal functions of the past and future tense markers, some linguists have proposed defining the non-present tenses not in terms of temporal precedence or sequence but rather of detachment, understood as either detachment from the present or detachment from reality: ‘non-actuality’ (Strang Reference Strang1968), ‘dissociation’ (Steele Reference Steele1975), or, in deictic terms, ‘distality’ (Langacker Reference Langacker1978) or ‘remoteness’ (Joos Reference Joos1964), as opposed to the ‘proximality’ of the present.

The present is what is before our eyes, hence (presumably) certain; what is detached from the present is uncertain to a greater or lesser extent. Already in 1956, Gonda commented on the epistemic values of the tenses (Gonda Reference Gonda1956).

But if there is some intimate connection between temporality and modality, just what is it? Jaszczolt proposes that our concept of time supervenes on modality, and in two ways: as supervenience of the concept of time on that of epistemic detachment, and as supervenience of the concept of time (‘internal’ time), intrinsically connected with evidentiality, on the properties of space/time (i.e., ‘real’ time).

While the author never fully explicates her use of ‘supervenes’ (a bit of philosophers' jargon likely unfamiliar to most linguists) – characteristically she says in Chapter 1 (‘Real time and the concept of time’) that ‘the type of supervenience will not be in focus’ (3) and in Chapter 2 (‘Time as modality’) that ‘the exact nature of [the] dependency [of epistemic on metaphysical modality] will not concern us here’ (39) – she speculates in several places (including 3, 39, 95, and 99), and in the end essentially proposes, that the relationship between temporality and modality is actually the stronger one of identity.

Thus her treatment is a reductive one: time, and hence tense, turns out just to be modality.

Jaszczolt argues that ‘internal’ time and evidentiality (which she sees as a version of epistemic modality) are intrinsically connected. And if tenses convey modality, she proposes that it is also the case that the representation of time in language is effected through the use of expressions whose semantics is thoroughly modal.

Not only is ‘internal’ time inherently modal, but so is ‘real’ time. It is not just the concept of time that is founded on the more primitive modal concept of detachment, but ‘real’ time itself is also founded on modality, namely metaphysical or ontological modality. It is this shared modality of ‘internal’ and ‘real’ time that enables the human concept of time to reflect the properties of the time of space–time (100).

Here a debatable assumption is required, namely that epistemic and metaphysical modality are both, for all intents and purposes, simply modality. This assumption is especially problematic as regards Jaszczolt's identification of epistemic modality with evidentiality. From the perspective of the semantics of human languages, it is hard to see just what the two have in common, given that epistemic modality, often marked with general expressions of modality, indicates the degree of possibility relative to a presupposed knowledge base, while evidentiality, generally marked in a distinctive manner, would seem to concern the degree of certainty of that base itself.

Whatever a philosopher might make of all this, a linguist must raise objections to the reduction, to the proposed modality of temporal semantics. The question is whether it follows from the functionality of language in the real world that the human concepts embodied in language, and hence the semantics of human languages, must in some way reflect the nature of reality, and whether in representing time, languages are in fact simply representing modality.

At the very least, the specification in this context of supervenience as identity is questionable on epistemological grounds: just what is it you believe when you believe that something is in the past, say? It surely is something more than that the situation in question is non-actual and uncertain. To be fair, this undoubtedly assumes a gross simplification (and possibly a distortion) of Jaszczolt's position, but it is quite likely what a linguist would come away with from the book.

The question of the correctness of the reduction arises especially in regard to Chapter 2, in the course of which, while trying to answer the main question of the essay (is time a primitive concept or do humans conceptualize time in terms of the more basic modal concepts of possibility and necessity?), Jaszczolt argues that temporal sequentiality is not real but can be traced to the underlying notion of modality in the form of epistemic possibility realized as degrees of acceptance of a proposition.

For this to work, evidentiality must simply be epistemic modality, and epistemic and metaphysic modality must both be, well, modality. To be sure, Jaszczolt makes a good argument for logical equivalence in each case, but where conceptualization is concerned, more than logic is at issue and it is not so clear that there is only one kind of modality, or that the difference or differences between epistemic and other types of modality are not relevant here. The linguist is tempted to protest that it is human, psychological conceptualization in question here, not the logical idealization that is embodied in Jaszczolt's model speaker and model addressee (as on 111).

While the discussion of conceptualist linguistics in this volume indicates that the goal here is to capture the way in which language represents time, we are told that ‘the ultimate objective is to build semantic representations of temporality and to demonstrate that for doing this modal concepts will suffice’ (39). They may suffice, but are they necessary? Can we not build adequate representations in which modal concepts are not utilized? This becomes an issue especially in regard to Chapter 4 (‘Time in Default Semantics’).

The author regards the analyses developed in Chapter 4 on the basis of the tenets of her theory of Default Semantics (an extension and modification of Discourse Representation Theory) as providing a validation of the theses set forth in the earlier chapters. We are told, in the conclusion to Chapter 4, that the representations presented in this chapter demonstrate the soundness of the arguments in Chapters 1–3 for the ‘modal basis of all temporality’ in that ‘the thesis that temporal location is construed as modal detachment can be succinctly stated in the form of semantic, and at the same time, mental, representations’ (164). But Jaszczolt's analyses seem merely to be compatible with the theories presented in the earlier chapters (as marked by her use of the word ‘can’ in this quotation) and while these analyses show that the theories possibly are correct, rather than their validating the earlier arguments, these analyses are themselves validated by those arguments. Consequently Chapter 4 seems more an attempt at the justification of Default Semantics than a justification of the reduction of tense to modality.

Regarding Jaszczolt's case, we must in the end return a Scotch verdict: not proven.