Evidence from signed languages
As we concluded, the first layers of grammatical evolution were restricted to lexical categories, which are the earliest forms of grammar that are accessible via grammaticalization in spoken (and written) languages. This raises the question of whether it may not be possible to push reconstruction back to linguistic forms that may have preceded lexical categories. There is no lack of works dealing with this question, nor of hypotheses that have been voiced; nevertheless, we will not deal with these hypotheses here because they are generally based on premises that are empirically not satisfactory in every respect.
But there is one possible exception and it concerns signed languages. As has been demonstrated in recent work, signed languages show roughly the same kinds of grammaticalizations that we observe in spoken languages (see e.g. Janzen 1995, 1998, 1999; Janzen and Shaffer 2002; Morford 2002; Sexton 1999; Pfau 2004; Pfau and Steinbach 2005, 2006; Shaffer 2000; Wilbur 1999). We saw above that a number of the pathways outlined for spoken languages have analogs in signed languages (see the seminal work of Pfau and Steinbach 2006 for a general treatment).
But there are also what appear to be modality-specific grammaticalizations in signed languages which are undocumented in spoken languages. For example, in German Sign Language (DGS), a grammaticalization from noun to auxiliary has been observed (Pfau and Steinbach 2006), and—to our knowledge—the development of the adjective WRONG of American Sign Language into an adverbial marker for unexpected events (Morford2002:331) also has no equivalent in spoken languages. However conceptually plausible these processes may be, we are not aware of any parallel of them in spoken or written languages (for an exceptionally interesting development of the noun ‘thing’ → ‘long object used as an instrument’→ purpose marker, see Epps2007).
Perhaps more importantly, research on signed languages may also shed light on the question posed above, namely on whether it may not be possible to push reconstruction back to linguistic forms that preceded lexical forms. There is little doubt that at least some lexical items of signed languages that were grammaticalized to functional categories have their source in gestures, that is, in communicative expressions using the hands and face. Thus, Janzen and Shaffer (2002) propose the reconstructions in (65); note further that Wilcox and Wilcox (1995) argue that there are gestural origins for markers of evidentiality (see also Pfau and Steinbach 2006).
But work on signed languages provides yet another perspective that is immediately relevant for the reconstruction of language evolution. Our discussion in “Layers” suggested that functional categories can be traced back ultimately to lexical categories. As it surfaces from recent research (especially Janzen 1999; Janzen and Shaffer 2002; Pfau and Steinbach 2005, 2006), however, there is an alternative pathway for functional categories, namely one that bypasses the lexicon, leading straight from manual or non-manual gesture to functional marker, such as the ones summarized in (66). In (66a) for example, the gesture proposed as the origin of the polar (= yes–no) question marker in both American Sign Language and British Sign Language (Woll 1981) is an eyebrow raise extending over the entire proposition being questioned, which became the obligatory sign for polar questions (usually along with a forward head tilt), that is, a marker for a grammatical function. Subsequently, this marker was further grammaticalized to a topic marker, first within the pragmatic domain, subsequently within the syntactic, and finally the textual domain.
On the basis of such observations one might conclude that ultimately lexical material might not be the only source for functional categories. Conceivably, such a hypothesis could be strengthened by more detailed analysis of suprasegmental grammatical forms as they show up, for example in tone or intonation contours of polar question marking. Unfortunately, the evidence available does not allow for any generalizations on this issue. But this raises more generally the question of what role gesture and prosody may have played in the evolution of spoken languages. Since we are restricted here to segmental speech, we have little to say about this issue except that it requires more research.
