Loanwords

The borrower may adopt the donor’s word along with the object or practice; the new form in the borrower’s speech is then a loanword.

The acquisition of a loanword constitutes in itself a lexical change, and probably we should say that it constitutes or entails a semantic change. A shape change is sometimes involved. [...] English acquisition of wiener [wí jnə r] involved no such change, since the language already had a morpheme represented by the shape [wí jn] and several morphemes represented by suffixed [-ə r]. Our acquisition of allegro [ə lé grow], on the other hand, entailed a shape change of the type just described.

Other kinds of phylogenetic change are not directly implied by a single new importation, but they may come about as the result of a whole wave of loanwords from some single source, along the following lines:

Grammatical change

ME acquired a large number of Norman-French adjectives containing the derivational suf­fix which is now - ablej/-ible: agreeable, excusable, variable, and others. At first, each of these whole words must have functioned in English as a single morpheme. But English has also borrowed some of the verbs which in French under­lay the adjectives, and in due time there came to be a large enough number of pairs of borrowed words for the recurrent termination to take on the function of a deriva­tional affix in English. This is shown by the subsequent use of the suffix with native English stems: bearable, eatable, drinkable (the stems tracing back to OE /bé ran/, /é tan/, /drí nkan/). [....]

...It is to be noted that the derivational affix was not borrowed as such: it occurred as an integral part of various whole words, and only the latter were actually borrowed. Apparently we can generalize on this point: loanwords are almost always free forms (words or phrases); bound forms are borrowed as such only with extreme rarity. [....]

Alternation Change

Our learned vocabulary, borrowed directly or indirectly from Latin and Greek, includes a good number of words like datum: data, phenomenon: phe­nomena, matrix: matrices. What has happened here is that we have borrowed both the singular and the plural forms of the word. [...] Since English already had the inflectional category of number, these importations do not imply any grammatical change, only additional patterns of alter­nation. In such cases there is usually competition between the imported and native patterns. Most of us tend to use data as a singular “mass-noun”, like milk, saying this data is rather than these data are. Doublet plurals in competi­tion are even commoner: matrixes [mé jtriksə z] and matrices [mé jtrə sijz], automata and automatons, gladioluses and gladioli. One cannot safely predict which alternative in such a case will in the end win out; currently, in English, the imported plural has a more learned connotation than the native one. […]

Phonemic and Phonetic Change

The first few members of a community to use a word from another language, or from a highly divergent dialect of their own, may imi­tate the pronunciation of the model accurately. Any iso­lated borrowing which spreads into general usage, however, is unlikely to retain its foreign pronunciation if that in any way goes against the pronunciation habits of the bor­rowers. […] Some of us pronounce initial [ts] in tsetse fly, tsar; most, however, begin the words with [z]. Even French words like rouge, garage, mirage, probably end more commonly in English with [j] than with [ž ].

However, it would seem that a great flood of loanwords from some single source, involving many bilinguals as the channel for the borrowings and with a major prestige fac­tor, can have some striking consequences in articulatory habits. The stock example, once again, is the influence of Norman French on English: it was through this influence that English acquired initial [v z j], and, consequently, the phonemic contrast between [v] and [f], [z] and [s]. [...]