Loanword Typology: why?

(a longer, but still preliminary position paper is available here, dated from June 2003)

How likely is it that a word with a given lexical meaning will be borrowed from one language into another language? At the moment, this question can be answered only on the basis of impressionistic observations, such as "body part terms are unlikely to be borrowed", or "terms for new artifacts are often borrowed". But it would be interesting to know more about the likelihood of borrowing for particular lexical meanings, for a variety of reasons:

(a) In assessing genealogical relatedness between languages, it is important to separate inherited material from borrowed material. Loanwords point to historical contact between two languages (there must have been speakers with sufficient knowledge of both languages at some stage), but not genealogical relatedness (i.e. descent from a common ancestral language). But which words are the most likely to be inherited? Linguists often assume that there is a set of words that are highly stable, unlikely to be replaced by borrowings, meaning shift, or new formations, but this notion needs to be made more precise to enhance the validity of conclusions based on it.

(b) The likelihood of lexical borrowing depends on the type of contact situation. A language of a population conquered by war may be likely to borrow military terms from the conquerors' language, and seafaring populations may contribute marine words to languages spoken inland. An invading population may borrow terms for local flora and fauna even if it is technologically and economically superior to the indigenous population. Once such generalizations have been securely established (on the basis of attested examples), it will be possible to draw inferences about the history of a population from the loanword patterns of a language.

(c) Borrowing patterns may be influenced by non-social, strictly linguistic factors, such as semantic complexity, abstractness, or taxonomic level, as well as syntactic factors such as word-class status (noun vs. verb vs. adjective; content word vs. function word). Generalizations in this area are interesting for linguistics-internal reasons because they may provide insights about the nature of language structure.

The idea of the proposed project is to study lexical borrowability using the classical methods of linguistic typology:
(i) establishing a world-wide sample of languages (ca. 30-40 languages)
(ii) surveying the types of loanwords found in these languages, on the basis of a fixed list of lexical meanings (ca. 1400 lexical meanings, based on the list used in the Intercontinental Dictionary Series)
(iii) attempting generalizations across the languages of the sample

In practical terms, prospects for success seem to be greatest if each language is in the responsibility of a single author, who is a specialist of the language and its history. The author would provide a list of all (known and suspected) loanwords among the 1400 lexical meanings of the fixed list, as well as information on what is known about the historical circumstances under which these words were borrowed. The author may of course discuss additional loanwords (beyond the 1400-word list), but the fixed list is important so that quantitative generalizations across the sample languages can be made.

The concrete outcome of this project will be an edited volume, consisting of the 30-40 chapters on the sample languages, plus an introduction and a generalizing chapter by the project leaders.