This article may require cleanup to meet Wikipedia's quality standards. (September 2010) (Learn how and when to remove this template message)
Language transfer (also known as L1 interference, linguistic interference, and crosslinguistic influence) refers to speakers or writers applying knowledge from one language to another language. It is the transfer of linguistic features between languages in the speech repertoire of a bilingual or multilingual individual, whether from first ("L1") to second ("L2"), second to first or many other relationships. It is most commonly discussed in the context of English language learning and teaching, but it can occur in any situation when someone does not have a native-level command of a language, as when translating into a second language.
When the relevant unit or structure of both languages is the same, linguistic interference can result in correct language production called positive transfer: here, the "correct" meaning is in line with most native speakers' notions of acceptability. An example is the use of cognates. However, language interference is most often discussed as a source of errors known as negative transfer, which occurs when speakers and writers transfer items and structures that are not the same in both languages.
Within the theory of contrastive analysis, the systematic study of a pair of languages with a view to identifying their structural differences and similarities, the greater the differences between the two languages, the more negative transfer can be expected. For example, in English, a preposition is used before a day of the week: "I'm going to the beach on Friday." In Spanish, instead of a preposition the definite article is used: "Voy a la playa el viernes." Novice Spanish students who are native English-speakers may produce a transfer error and use a preposition when it is not necessary because of their reliance on English. According to Whitley, it is natural for students to make such errors based on how the English words are used. Another typical example of negative transfer concerns German students trying to learn English. Since the German noun "Information" can also be used in the plural - "Informationen" - German students will almost invariably use "informations" in English, too, which is grammatically wrong. From a more general standpoint, Brown mentions "all new learning involves transfer based on previous learning". That could also explain why initial learning of L1 will impact the learning of L2.
The results of positive transfer go largely unnoticed and so are less often discussed. Nonetheless, such results can have a large effect. Generally speaking, the more similar the two languages are and the more the learner is aware of the relation between them, the more positive transfer will occur. For example, an Anglophone learner of German may correctly guess an item of German vocabulary from its English counterpart, but word order, connotations and collocation are more likely to differ. Such an approach has the disadvantage of making the learner more subject to the influence of "false friends".
In addition to positive transfer resulting in correct language production and negative transfer resulting in errors, there is some evidence that transfer from the first language can result in a kind of technical, or analytical, advantage over native (monolingual) speakers of a language. For example, second-language speakers of English whose first language is Korean have been found to be more accurate with perception of unreleased stops in English than native English speakers who are functionally monolingual because of the different status of unreleased stops in Korean from English. That "native-language transfer benefit" appears to depend on an alignment of properties in the first and the second languages that favors the linguistic biases of the first language.
During the 1950s, memory research began investigating interference theory. This refers to the idea that forgetting occurs because the recall of certain items interferes with the recall of other items. Throughout the 1950s, researchers provided some of the earliest evidence that the prior existence of old memories makes it harder to recall newer memories and he dubbed this effect "proactive interference". During the same time, researchers began investigating negative transfer. Negative transfer concerns itself with a detrimental effect of prior experience on the learning of a new task, whereas proactive interference relates to a negative effect of prior interference on the recall of a second task.
The most obvious and used proactive interference and negative transfer paradigm from the 1950s and 1960s was the use of AB-AC, or AB-DE lists. Participants would be asked to learn a list of paired associates in which each pair consists of a three letter consonant-vowel-consonant, nonsense syllable (e.g. DYL), used because it was easy to learn and lacked pre-learned cognitive associations, and a common word (e.g. road). In this paradigm two lists of paired associations are learned. The first list, (commonly known as the AB list) would consist of nonsense syllables as a primer (which constituted the 'A' term), followed by a word (which constituted the 'B' term). The second list would consist of either the same nonsense syllable primer and a different word (A-C list) or a different nonsense syllable primer and a different word (D-E list). The AB-AC list was used because its second set of associations (A-C) constitutes a modification of the first set of associations (A-B), whereas the AB-DE list were used as a control.
Shortly afterwards proactive interference was demonstrated with the Brown-Peterson paradigm. A single Brown-Peterson trial consists of a study list, a retention interval and then a recall period. Each list may consist of a handful of related items and are presented individually every few seconds. For the duration of a short retention interval, subjects are then asked to perform an engaging distractor task such as counting backwards in sevens, or thinking of an animal with every letter in the alphabet to minimize rehearsal. Subjects are then asked to recall the items from this second list. Although the lists from previous trials are now irrelevant, the fact that they were studied at all makes it difficult for subjects to recall the most recent list.
Negative transfer was examined by researchers in the 60s and found differential learning between trials. Specifically, differences in the learning rates of list 2 provided clear evidence of the negative transfer phenomenon. Subjects learned an A-C paired association list to a criterion of all associations correct, following learning a list of A-B paired associations to criterion. Ultimately, it was found that those subjects took an increased number of trials to complete the learning task compared to subjects who did not learn the A-B list or from subjects who had to learn a D-E list.
Language transfer may be conscious or unconscious. Consciously, learners or unskilled translators may sometimes guess when producing speech or text in a second language because they have not learned or have forgotten its proper usage. Unconsciously, they may not realize that the structures and internal rules of the languages in question are different. Such users could also be aware of both the structures and internal rules, yet be insufficiently skilled to put them into practice, and consequently often fall back on their first language. The unconscious aspect to language transfer can be demonstrated in the case of the so-called "transfer-to-nowhere" principle put forward by Eric Kellerman, which addressed language based on its conceptual organization instead of its syntactic features. Here, language determines how the speaker conceptualizes experience, with the principle describing the process as an unconscious assumption that is subject to between-language variation. Kellerman explained that it is difficult for learners to acquire the construal patterns of a new language because "learners may not look for the perspectives peculiar to the [target/L2] language; instead they may seek the linguistic tools which will permit them to maintain their L1 perspective."
The conscious transfer of language, on the other hand, can be illustrated in the principle developed by Roger Andersen called "transfer-to-somewhere," which holds that "a language structure will be susceptible to transfer only if it is compatible with natural acquisitional principles or is perceived to have similar counterpart (a somewhere to transfer to) in the recipient language." This is interpreted as a heuristic designed to make sense of the target language input by assuming a form of awareness on the part of the learner to map L1 onto the L2. An analogy that can describe the differences between the Kellerman's and Anderson's principles is that the former is concerned with the conceptualization that fuels the drive towards discovering the means of linguistic expression whereas Andersen's focused on the acquisition of those means.
Transfer can also occur in polyglot individuals when comprehending verbal utterances or written language. For instance, German and English both have relative clauses with a noun-noun-verb (=NNV) order but which are interpreted differently in both languages:
German example: Das Mädchen, das die Frau küsst, ist blond
If translated word for word with word order maintained, this German relative clause is equivalent to
English example: The girl that (or whom) the woman is kissing is blonde.
The German and the English examples differ in that in German the subject role can be taken by das Mädchen (the girl) or die Frau (the woman) while in the English example only the second noun phrase (the woman) can be the subject. In short, because German singular feminine and neuter articles exhibit the same inflected form for the accusative as for the nominative case, the German example is syntactically ambiguous in that either the girl or the woman may be doing the kissing. In the English example, both word-order rules and the test of substituting a relative pronoun with different nominative and accusative case markings (e.g., whom/who*) reveal that only the woman can be doing the kissing.
The ambiguity of the German NNV relative clause structure becomes obvious in cases where the assignment of subject and object role is disambiguated. This can be because of case marking if one of the nouns is grammatically male as in Der Mann, den die Frau küsst... (The man that the woman is kissing...) vs. Der Mann, der die Frau küsst (The man that is kissing the woman...) because in German the male definite article marks the accusative case. The syntactic ambiguity of the German example also becomes obvious in the case of semantic disambiguation. For instance in Das Eis, das die Frau isst... (The ice cream that the woman is eating...) and Die Frau, die das Eis isst... (The woman that is eating the ice cream...) only das Eis (ice cream) is a plausible object.
Because in English relative clauses with a noun-noun-verb structure (as in the example above) the first noun can only be the object, native speakers of English who speak German as a second language are likelier to interpret ambiguous German NNV relative clauses as object relative clauses (= object-subject-verb order) than German native speakers who prefer an interpretation in which the first noun phrase is the subject (subject-object-verb order). This is because they have transferred their parsing preference from their first language English to their second language German.
These examples could be multiplied endlessly to reflect the linguistic interactions of speakers of the thousands of existing or extinct languages.
With sustained or intense contact between native and non-native speakers, the results of language transfer in the non-native speakers can extend to and affect the speech production of the native-speaking community. For example, in North America, speakers of English whose first language is Spanish or French may have a certain influence on native English speakers' use of language when the native speakers are in the minority. Locations where this phenomenon occurs frequently include Québec, Canada, and predominantly Spanish-speaking regions in the US. For details on the latter, see the map of the Hispanophone world and the list of U.S. communities with Hispanic majority populations. The process of translation can also lead to the so-called hybrid text, which is the mixing of language either at the level of linguistic codes or at the level of cultural or historical references.