Context, Compositionality and Amity: A Response to Rett

Introduction

In an insightful and provocative paper, Jessica Rett (2006) claims that attempts to locate the (non-indexical, non-demonstrative) semantic contributions of context in syntax run into problems respecting compositionality. This is an especially biting problem for hidden indexical theorists such as Stanley (2000, 2002) who deploy hidden variables to provide a compositional theory of semantic interpretation. Fortunately for the hidden indexical theorists, her attack fails, albeit in interesting and subtle ways.

The following paper is divided into four sections. Section I presents a skeletal version of Rett’s argument. Those already familiar with Rett (2006) can skip ahead without shame. Section II offers a defense to the hidden indexical theorists. The defense will involve distinguishing the determinants of sentence meaning relative to a context from the information a hearer uses to figure out what the sentence meant in the context. Rett’s attacks concern the latter, but hidden indexical theorists tend to be concerned with the former. Clarity on this issue affords a better view of the moral to be drawn from Rett’s considerations. Section III offers a further argument against Rett: her argument applies equally well to visible pronouns. This shows that her argument, if sound, is of concern to everyone and not just hidden indexical theorists. I’ll conclude in section IV with a general diagnosis of Rett’s argument.

I Rett’s Argument

It is uncontroversial that most utterances of any of the following sentences are correctly interpreted as expressing a richer meaning than the sentences seem to express:

1. 1. It’s raining.

2. 2. Every city suffered from the bombing.

3. 3. Nathan Explosion is tall.²

Utterances of (1) tend to be correctly interpreted as claims about the weather in some salient location, (2) as expressing quantification over some restricted domain and (3) as attributing to Nathan the property of being tall given some contextually determined scale and a point on that scale. A general diagnostic (though by no means necessary or sufficient condition) for such sensitivity involves the following test: are there pairs of contexts of utterance where utterances of the same sentence is true relative to the first, false relative to the second in the same circumstance of evaluation. The answer here is clearly ‘yes’. (1) can be uttered (simultaneously) by Toki in Norway where it is raining and by Brendan in Springfield where it isn’t (relative to the same circumstances of evaluation).³ Similar thought experiments confirm similar results for (2) and (3).

It is a controversial claim, however, that the utterances do express more than the sentences express relative to those same contexts. If the interpretations are identical (despite appearances), a further interesting question involves what semantic and syntactic mechanisms allow for interaction with context in producing the appropriate interpretations. Many answers to these questions have been proposed. I will paint with broad strokes so as to not get bogged down in minutiae.⁴

Unarticulated Constituents (UC): UC theorists claim that semantic interpretation of the syntactic structures associated with (1)—(3) underdetermines what is said by utterers of the sentences and that pragmatics is called upon to ‘fill in the gap’ between semantic interpretation and meaning. In other words, utterance meaning is richer than sentence-relative-to-a-context meaning.⁵

Context Sensitivity (CS): CS theorists claim that the intensions of words such as ‘raining’ and ‘tall’ shifts depending on context, just as ‘I’ and ‘here’ do. Typically CS theorists posit that the relevant words have a complex character that is a non-constant function from context to contents.

Hidden Indexicals (HI): HI theorists (HITs) posit unpronounced variables that are part of the input to semantic interpretation—a sentence’s logical form, or LF. These variables act much like regular pronouns: they can be bound, or can take values from context and thereby provide the relevant interpretation of sentences such as (1)-(3) in context.

Rett’s official target is the HITs, with Stanley (2000, 2002) serving as a paradigm. HITs (about sentences like (1)) claim that (1)’s LF is augmented by a hidden variable(s) which has as its...