Losing English as a First Language

Abstract

LOSING ONE'S FIRST LANGUAGE (LANguage attrition) has been documented in a wide variety of settings. Although some anecdotal evidence exists of language loss in situations of extreme isolation from all language contact, such as arctic explorers, most documented cases of language loss involve language contact or the learning of a second language. Language loss is a broad term which can include complete loss, such as language death (10; 11; 12; 13), loss of proficiency (27; 28; 49; 50; 52), and various forms of modification in language contact situations (4; 7; 8; 18; 57; 58; 60; 62). These latter cases (modification of a first language) are sometimes excluded from the term language loss because speakers may continue to be completely fluent in their first language. However, in another sense these fluent speakers should be considered cases of language loss because fluency and proficiency are not synonymous. One may speak a language fluently but still exhibit forms which deviate substantially from those of native speakers. Therefore, in order to avoid confusion, the term first language loss refers to all forms of loss, complete or partial (modification). Although first language loss in its various forms has been studied systematically for some time (62), there seems to be no study that involves stylistic variation in L1 loss, that is, L1 variation taking into account the influence of L2. This fact is remarkable given that there are numerous studies of both L1 variation (16; 32; 35; 43; 58) and L2 variation (2; 3; 9; 40; 42; 44; 45; 48; 53; 56). However, the general trends found in L1 and L2 variation are perhaps indicative of what one would expect to find in L1 loss. The general trends in L1 and L2 variation studies are similar.' Both L1 and L2 studies typically show that speakers approximate the perceived standard with greater accuracy as the style becomes more formal. An early study by Fischer concerning the pronunciation of -ing by New England children (ages three to ten) showed that girls used [rj] more frequently than boys, but in both sexes the [Uj] variant was used more frequently as style became more formal. N merous studies of postvocalic r, exemplified by some of the early work of Labov (33; 35), reveal postvocalic r is correlated with both social class and formality. L2 studies show similarities in that greater target language accuracy occurs in more formal styles (9; 22; 47).2 For example, the Dickerson and Dickerson study found that Japanese learners of English produced English /r/ with nearly 100% accuracy in word lists but with only fifty percent accuracy in conversation. The similarities of L1 and L2 studies may best be explained on the basis of how closely speakers monitor their speech: the more they monitor the greater the accuracy. The term monitor, or the amount of attention one pays to speech, originally used with reference to L1 variation by Labov (34), was later used with a si ilar definition for L2 studies (2; 54; 55; 56).3 With these generalizations of L1 and L2 variation in mind, what would one expect of L1 variation for speakers who are under the strong influence of L2? Because by definition the vernacular or casual represents the most natural, unmonitored style, one might anticipate this style to be the least disturbed by L2 influence. On the other hand, it is reasonable to expect the opposite, considering the general concept of interference and under what situations it is