There is an ongoing debate on whether protolanguage was based on expressions akin to individual words or on expressions akin to whole sentences. The first view is usually called synthetic or compositional and the second view holophrastic. In this talk I will draw attention to the ways in which philosophy of language might contribute to this debate, and in particular to the discussion in metasemantics (i.e., the metaphysical inquiry about semantic facts and properties) about the fundamental bearers of linguistic meaning. On one view, the fundamental bearers of meaning are whole sentences, and word meaning is derivative. Call this view Sentence Priority. Defenders of Sentence Priority include Ludwig Wittgenstein, W. V. O. Quine, Jeremy Bentham, Donald Davidson and Paul Grice, just to name a few. On another view, the fundamental bearers of meaning are individual words. This is sometimes called the building block view, and has not been very popular historically. The parallel between this discussion in metasemantics and the compositionality vs. holophrasis debate in evolutionary linguistics is striking. If the comparison between the two can be substantiated, then one discussion might illuminate the other in interesting ways. For example, defenders of Sentence Priority argue that there are no facts that can determine or fix the meaning of individual words apart from facts that determine or fix the meaning of whole sentences. This entails that individual words cannot be endowed with meaning in isolation: whatever meaning they have is derived from the meaning of the whole sentences in which they occur. In other terms, the meaning of whole sentences is metaphysically prior to the meaning of individual words. If this is right, then it seems that protolanguage has to be holophrastic, as a matter of metaphysical necessity: whatever facts determined the meaning of our ancestor’s expressions, they must be facts involving whole messages and their adequate protolinguistic vehicles; this, in turn, is reason to believe that these vehicles were more like sentences than individual words. On the other hand, there are many epistemological reasons for thinking that protolanguage expressions were word-like rather than sentence-like (e.g., it is difficult to agree on the meaning of holophrastic utterances; holophrastic signals are not stable enough across time). If the comparison makes sense, similar arguments can be leveled against Sentence Priority in metasemantics.
In short, then, my goal is to explore this apparent connection between metasemantic debates in philosophy of language and debates in evolutionary linguistics, in the hope that one might learn from the other.