About the forms of primitives

15 W and her colleagues have recently added an additional requirement to Leibniz’s program: a primitive must prove itself cross-linguistically. That is, these primitives must have exact lexical equivalents in a wide variety of unrelated languages. The potential primitives world and imagine were dropped from NSM because they were not found in a number of languages.

2.4 About the forms of primitives

An important part of establishing primitives cross-linguistically is learning how to recognize them. The first step in this process is understanding the ways in which a primitive’s form may differ across languages. The following list addresses a number of common misconceptions: 1. A primitive may appear to be formally complex. Something and someone appear to consist formally of some + thing and some + one. Likewise, Japanese nanika and dareka appear to contain a formative -ka. Semantically, however, some + one does not equal someone. Someone is a simple and irreducible concept. 2. A primitive may be composed of several possibly noncontiguous wordsmorphemes. The primitives THERE IS, IS SOMEWHERE, IF ... WOULD, and KIND OF are realized in SE as several words. 3. A primitive form may be polysemous, with both a primitive and a non-primitive sense. Know is polysemous between a non-primitive sense as in we know John and the primitive sense as in we know John is a good boy. In Yunkuntjara Goddard 1994, the form for the primitive WANT can also mean stomach. 4. A primitive may have contextual variants allolexes such as I and me in English. 5. A primitive need not be realized as a free form, it may be realized as a bound-morpheme. One exponent for WANT in Japanese is the verbal suffix -tai. 6. A primitive in one language need not be in the same syntactic category as its English counterpart.The other exponent for WANT in Japanese is the adjectival hoshii. Neither -tai nor hoshii are verbs like English want. From NSM’s point of view, the primitive’s formal syntactic category is not as important as its semantic combinability. It does not matter whether NSM substantives such as I or THIS are realized as nouns or verbal inflections or whether NSM predicates such as WANT are realized as verbs or adjectives. What does matter is that there is a mechanism by which they may be combined to form a canonical sentence meaning I WANT THIS. Combinatorial operations are left unspecified in NSM and they may be accomplished through any means, whether syntactic or morphological. 7. Under the current strong lexicalization hypothesis, a primitive must have some kind of segmental content. It cannot be realized as an intonation or as a grammatical process such as reduplication. This constraint may seem too strong to some, but for the time being, it is retained because it is the strongest and therefore the most interesting hypothesis.

2.5 Cross-linguistic equivalence