Reconstructions

Note

The most challenging feature of this comparison is its meaning. Dempwolff (1934-38) posited *qutan ‘Wald’ ( = ‘woods, forest’). However, reflexes of PMP *halas almost invariably mean ‘forest’, while reflexes of PMP *qutan cover a wide range of referents, including at least ‘vegetable’ (WMP + CMP), ‘weed’ (WMP + OC), ‘forest’ (WMP + CMP + OC), ‘wild’ (WMP + CMP) and bean (WMP + CMP). This striking difference in "semantic profile" (Blust 1987) clearly favors the reconstruction of PMP *alas with the meaning ‘forest’, but it does not mechanically provide us with a gloss for PMP *qutan. To gloss the latter we must consider both the range of its semantic reflexes and its relationship to other PMP terms, such as *alas.

At first sight there is little apparent connection between such meanings as ‘vegetable’ and ‘forest’, and we might therefore treat the reflexes of *qutan which cluster about each of these meanings as belonging to different cognate sets. Apart from the problems of synonymy already mentioned in connection with *alas, what makes it difficult to segregate these semantic clusters under two homophonous PMP reconstructions, *qutan_₁ and *_qutan_₂, is the presence of forms which provide a clear semantic bridge between them. First, the meanings ‘vegetable’ and ‘wild’ can be reconciled through a hypothesis that PMP *_qutan referred to wild vegetables, a meaning that is actually reflected in Alune utane. Second, the semantic reflex ‘weed’ (and the similar meaning ‘grass’ in Maranao and Palauan) can be explained by modifying our hypothesis to state that *qutan referred to small, wild herbaceous plants whether they were edible/useful or not. Third, the meaning ‘bean’ can be included as an implicit component of our reconstructed gloss by hypothesizing that leguminous plants constituted an important part of the total inventory of small, wild herbaceous plants known to speakers of PMP. Finally, the meaning ‘forest’ can be added under the assumption that *qutan referred both to a type of vegetation and to the type of land on which it grew.

The difficulty of glossing PAn *quCaN stems from the paucity of known Formosan reflexes. For PMP *qutan this problem does not exist; rather, it is difficult to arrive at an adequate gloss because there appears to be no single convenient English lexical category which corresponds to the PMP concept. If we gloss *qutan as ‘small, wild herbaceous plants; scrub-land, bush’ we will perhaps come close to the sense of the PMP term. A meaning very similar to the first half of this gloss is reflected in Ngadha uta ‘underbrush, scrub, low-growing plants’ and Lakalai la-huta-huta ‘general term for small plants and leaves’. Apart from its additional reference to type of land the meaning of PMP *qutan is essentially identical to what Brown (1984:13) has called grerb: "Small plant (relative to the plant inventory of a particular environment) whose parts are chiefly herbaceous (green, leafy, nonwoody)."

The basic correctness of this interpretation is strikingly confirmed by a secondary meaning of Lakalai la-huta-huta, ‘trash’, since Brown (1984:62) has argued that "terms for grerb in both Polynesian and Mayan languages have developed from words for rubbish, garbage, trash, litter, rotten stuff, and the like. In most cases such words were first extended to weeds, i.e., useless and bothersome small herbaceous plants, and then through expansion of reference to small herbaceous plants in general." However, in the present case there can be no doubt that the semantic development in Lakalai la-huta-huta has been from ‘grerb’ to ‘trash’ rather than the reverse. Finally, it is notable that in Southeast Ambrym the reflex of PMP *qutan acquired a sense very similar to that associated with reflexes of PMP *banua ‘inhabited territory’ in other OC languages, and like some of these came to mean ‘weather’, thus providing another example of what I have elsewhere called a ‘semantic cycle’ in historical change (Blust 1989b).