Chapter VI, Part 1

VI. Interlingual Discourse in Theory and Practice

1. The whole range of communicative events activating more than one language can be termed interlingual discourse. At one end of the range is ‘nonce borrowing’, where an individual text producer occasionally throws in just a foreign word or two without following or setting any precedent. At the other end is the fully stabilized ‘creole’, where a whole community uses a language system that was originally derived by mixing two or more languages. In between lies an extensive gradation of personal and social combinations. Here again is a widespread practice where theory has lagged far behind. For a °mainstream linguistics studying each language by itself° as ‘a system wherein everything holds everything else in place’ (II.44), interlingual discourse would seem merely ‘unsystematic’. In contrast, our own conception of °discourse being an actual system ° could be expanded for discourse that draws upon more than one °virtual system  of language° while °interfacing them with multicultural cognitive and social constraints°.

2. Interlingual discourse also raises the troublesome issue of how to distinguish between language versus dialect. Following upon the philological study of dialects (II.16f, 21), early linguists were perplexed by the prospect that ‘language’ and ‘dialect’ might be ‘purely relative terms’ (Sapir). Somewhat deviously, they argued that the ‘geographical diversity’ among dialects is really ‘temporal diversity’ affecting a single language (Saussure) and could thus be excluded from the °synchronic perspective outside of time°. Dialects could only be produced by the break-up of a single genuine language: ‘all languages that are known to be genetically related’ would be merely ‘divergent forms of a single prototype’ (Sapir). This general formulation by no means accounts for the available evidence over extended periods of historical change, e.g. when different languages grow more similar instead of less so. Nor does it indicate why and how such a break-up might occur and what could result when the dialects remain in contact, even when they have become different ‘languages’ by the usual criteria. The chief missing link here, as in any linguistics of °language by itself°, is precisely interlingual discourse, which must be the main social practice for navigating evolution toward either unity or diversity. There, real speakers easily overstep the tidy boundaries that linguistics would prefer to draw around the individual language or dialect.

3. Understandably, interlingual discourse also raises problems about sources of evidence. Indirect evidence might be gleaned by comparing and contrasting languages, and inquiring whether and how they may at one time have been used by the same groups in the same discourse. Here, we can seldom draw firm conclusions, because various factors may be influencing the data: language evolution, historical relatedness, migrations, language contacts, social status, colonialism, or just the frozen accidents occurring independently in several languages. We can seldom isolate a single factor that enabled a given feature or pattern to spread from one language into another. More direct evidence can be found when items clearly belonging to different languages are observed appearing together. Here, we can record some useful data and consult the participants as informants. Yet the motives or reasons for the phenomena may still be complex. Items or patterns may have entered unintentionally by interference, intentionally by borrowing, or spontaneously by code-switching. Much work remains to be done for theory to catch up with practice.

VI.A. Indirect evidence: Comparing and contrasting languages

4. Research on more than one language can either compare them by foregrounding their similarities or else contrast them by foregrounding their differences. Comparing was naturally the central project of the domain of comparative philology, which helped to prepare the foundations for the science of linguistics (II.16-20). Ingenious methods were developed to uncover subtle similarities among languages and to determine whether a common ancestry or history might be involved. Languages with rich points of comparison were widely held to be members of the same language family, e.g., the Slavic languages. The notion that ‘language developed but once in the history of the human race’ (Sapir 154) has had an ancient tradition and has recently resurged in studies of ‘monogenesis’ (cited for II.19).

5. Yet assuming that all languages descended from just one might make it harder to explain the origin of languages as an event requiring special or even unique conditions, and still more so to explain the vast and diverse processes of evolution and change for creating the stunning diversity among all the known languages in the world. If, as suggested above, °natural language° emerged by ‘externalizing’ the already ‘amplified’ internal languages of human life-systems (III.106ff), several scenarios are open: (a) the origin occurred only once, and the species that developed it survived the others; (b) the origin occurred more than once, but only one species survived; (c) the origin occurred more than once, each time giving rise to the ancestor of one language family that evolved down into historical times. Though we have scant evidence for deciding which scenario actually occurred, the question is highly significant for theories of language change and evolution. Scenarios (a) and (b) imply much more radical processes for change than does (c); and (b) raises the salient question of which language type could favor survival of the species over other types.

6. A geographical and social source of evolution and change has been explored under the term languages in contact. This is a common practical situation, most pronounced after individuals or groups have migrated from their home localities, e.g., to engage in commerce or warfare. Whether one culture chose either to segregate itself from another or else to integrate with it, languages would occupy a primary role as the most elaborated signifiers of cultural identity. According to the Christian Bible, for example, the word ‘shibboleth’ (‘ear of grain’ in Hebrew) served to segregate the conquering people of Gilead from the defeated people of Ephraim, who pronounced it ‘sibboleth’ (cf. Book of Judges 12: 4-6). Even today, when language contacts have been vastly intensified by regional mobility and multiculturalism, the presence of people speaking an unfamiliar language can cause irritation or discomfort and foster an impulse to be separated. Such feelings are avidly exploited by the militant monoculturalism portrayed in VII.B, leading to a host of confrontations in everyday life and work.

7. To be an effective cause of language change, language contacts would need to influence a substantial group of speakers and their discoursal practices. The effects should be strongest when widespread discourse regularly exposes speakers to multiple influences. Unfortunately, such discourse leaves behind little historical evidence or documentation, especially among languages that have disappeared or have not developed writing systems. Most written texts from earlier historical periods belonged to °prestigious text types° (II.3ff), whose producers might have considered it quite inappropriate to mix languages. On the contrary, most °language guardians° regard the mixing of languages with amusement, ridicule, or downright alarm, witness their snobbish disrespect for pidgins and creoles (cf. VI.50, 52).

8. The prospective role of language contacts as a cause of language change thus remains in dispute. Sometimes we have considerable evidence for the influence of one language upon the evolution of another, e.g., when English was overlaid by Norman French as a language of administration from the 11th century onward, or when ‘ancientist’ (‘classical’) French of the 17th century was expressly modeled on literary Latin. But even then it’s hard to determine why such an influence had the specific effects it did. The evidence is most straightforward for lexical borrowing, e.g., when English borrowed French names of certain animals to designate their edible meat, whence the present-day resemblance of ‘beef’ to ‘boeuf’, ‘veal’ to ‘veau’, ‘mutton’ to ‘mouton’, and ‘pork’ to ‘porc’ (cf. III.124; VII.152). Who actually did the borrowing and how it came to be accepted by the language community is usually not recorded; all we see are the attestations. Nor do we have adequate samplings of transitional discourses where the changes are still in the process of becoming accepted.

9. Our gravest problem is to determine how language contact and interlingual discourse might contribute to sweeping changes of an entire subsystem, such as phonology, morphology, or grammar. Grammatical borrowing is far harder to explain than lexical. The evidence is most secure when a variety of a language differs from the more ‘standard’ variety precisely in ways that make it more similar to the languages it contacts. In the village of Kupwar in southern India, four languages have been spoken for some 400 years: the Indo-Iranian languages Urdu and Marathi and the unrelated Dravidian languages Kannadu and Telugu. The long contact is probably why the local varieties of these languages show some conspicuous formal similarities. Whereas in standard Urdu the two Genders are mainly grammatical but conceptual for animate beings, Kupwar Urdu resembles Kannada in reserving Feminine for female humans only, while non-human animates are all Masculine. Conversely, Kupwar Kannada resembles Urdu in having an expression for Representive ‘be’ in patterns like ‘Noun + “is” + Noun’, while standard Kannada has none.

10. Such interlingual contacts should best exert their effects when the languages are rather similar, e.g., the ‘tone languages’ of Southeast Asia like Vietnamese, Thai, and regional varieties of Chinese. But, due to the °non-deterministic nature of self-organization°, we cannot predict just which changes will actually appear or just how much interlingual discourse will be needed to set them in motion and see them through. Even with intense language contacts, it’s not clear why speakers would consciously borrow formal features or patterns or why other speakers with lesser contacts would decide to imitate them. Presumably, most changes occur gradually and without conscious effort or consent; yet the long-range effects of the changes can attain remarkable proportions. Mbugu (or Ma’a), a non-Bantu language spoken in and around Usambara in Tanzania, seems to have borrowed its complex Nominal and Verbal morphology wholesale from the Bantu family, even though most of its Nominal and Verbal Stems as well as all Pro-Nouns resemble those of the Cushitic languages usually assigned to the Hamito-Semitic family. Such a shift could scarcely occur without long-range interlingual discourse among substantial groups of speakers — and even so it baffles explanation. Similar contacts would be needed if the Slavic languages, in contrast to their unaffected Baltic relatives, changed their Vowel systems through contact with Finno-Ugric speakers, as Herbert Galton suggests.

11. In formal linguistics, comparing languages has been less influential than contrasting them. Fieldwork made important progress by describing remote languages whose organization contrasted dramatically with the familiar Center languages, and by attaining a heightened sensitivity for possible modes of organization (cf. II.31ff, 37, 84ff, 109). Later on, research was inaugurated under such explicit headings as contrastive linguistics and interlinguistics for examining how languages differ. One well-known contrast between English and German, mainly expounded by Günther Rohdenburg, is the tendency of English to format Circumstances as Agents, e.g., ‘the theatre seats 1200 people’ (sample [81] in IV.50), which would not occur in German, e.g., ‘*das Theater setzt 1.200 Menschen’. Such work can have important applications to interlingual discourse in domains like language learning or translation (VI.14).

12. The preference for contrastive over comparative reflects the °formalist leanings of mainstream linguistics°: viewing languages as formal systems accentuates their differences and erases cultural memory for historical contacts. The similarities gain prominence when we view languages as functional systems whose °lexicogrammars interact with cognitive and social constraints°. For example, several European languages like German and Spanish hold a view of Agent and Initiative that favors Reflexives for a wide range of Actions such as Enactives like sitting down or lying down or Emotives like getting mad or being happy, whereas English favors Medials wherein the Agent is the Medium of the Action (cf. IV.49, 105) — even though German and Spanish are less closely related than are German and English. If, as Michael Halliday says, ‘the grammar of every natural language is (among other things) a theory of human experiences’, then shared experiences on the European continent might be expected to encourage similar designs, including peculiarities in formal details, e.g., Reflexives even for Actions that hardly seem self-directed, as in Spanish ‘míreselo!’ and German ‘schau dir das an!’ (versus English ‘look at that!’).

13. Conversely, lexicogrammars might differ to accommodate divergent views of human experience. For Ruqaiya Hasan, everyday spoken Urdu  contrasts with English by deploying ‘implicit devices’ whereby Participants in a Process need not be made explicit in discourse, because they are clear in the context of situation and are taken for granted within the culture (VII.69); for example, we see a higher proportion of unmarked ellipsis of the Clause Subject than we find in English (cf. IV.195). Or, for Jim Martin, Tagalog  has three main control centers for organizing its grammar: family, i.e., doing things together or sharing things; face, i.e., keeping up appearances and respecting other people’s status; and fate, i.e., events beyond the individual’s control; he contrasts these with the corresponding English control centers of individuality, reality, and determinism. Such tendencies recall the much-discussed ‘principle of linguistic relativity’, holding that ‘language is a guide to “social reality”’ to the degree that ‘the “real world” is to a large extent built upon the language habits of the group’ (Sapir); and that ‘the structure of a human being’s language influences the manner in which he understands reality and behaves with respect to it’ (Whorf). In the framework of this book, we might prefer to say that language, cognition, and social reality dialectically design each other, though the effects may vary widely in their directness, timing, consistency, thoroughness, and so on. We can expect to witness some dramatic effects as non-Western cultures at the Periphery, such as in the Philippines, get relentlessly °Westernized in the name of the economic growth° that mainly benefits the multinational corporations, the World Bank, the International Monetary Fund, and so on. The Center sends out ‘culture’ and draws in wealth (cf. VIII.33, 54ff).

14. Contrasts between lexicogrammars are also the concern of comparative stylistics, which explores the characteristic distributions of unmarked choices in two or more language systems. Incidental examples are easy to find, e.g., the unmarked tag-like follow-up ‘ouviu?’ in spoken Brazilian Portuguese where the English follow-up ‘did you hear me?’ would be marked as impatient or authoritarian. But comprehensive surveys are hard to find. One classic work is Jean Paul Vinay and Jean Darbelnet’s study of French and English. Toward the lexical end of the lexicogrammar, the study noticed ‘lexical modulation’ in how events and their participants are typically organized. A term may focus on Agent plus main Target in English versus on Agent plus enabling Target in French (e.g., ‘fireman’ versus ‘pompier’, an operator of water pumps); or on the Action versus the enablement (e.g., ‘air lift’ versus ‘pont aérien’, aerial bridge); or on the Action versus the Circumstance (e.g., ‘a first-run movie’ versus ‘un film en exclusivité’). But we also find terms where French focuses more on the Action than does English (e.g., ‘bouteille isolante’ [isolating bottle] versus ‘vacuum bottle’, or ‘papier peint’ [painted paper] versus ‘wall paper’). Toward the grammatical end, the study noticed different patterns of ellipsis (cf. IV.186) with French being more inclined to retain Substitutes (in the sense of IV.187-95), e.g.: ‘il pourrai le savoir’ versus ‘he might know’; ‘essayez de m’en empêcher’ versus ‘try and stop me’; ‘il est satisfait mais je ne le suis pas’ versus ‘he is satisfied but I am not’. Yet here too, the reverse also occurs, e.g., ‘finissez!’ versus ‘stop it!’, or ‘j’ai du mal á croire que…’ versus ‘I find it hard to believe that…’. Such comparisons again raise the interesting but difficult prospect of stating some sparse but consistent functional correlations between pairs of languages, such as would be extremely useful for teaching the language and for training translators (VI.12). We would need a deal of checking against the large multi-language data corpuses recently becoming available, such as the trilingual English-French-Spanish corpus at the University of Lancaster.

15. More specific comparisons might be drawn for particular styles, such as a ‘poetic’ style in French versus a ‘poetic’ style in English. Yet our general statements might still be fairly sparse; richer ones would need to address the more specific ‘poetic’ styles of historical periods, ‘schools’, or even individual poets. We might thus compare the influential style developed by Charles Baudelaire for 19th-century French poetry with the style of various English versions, e.g., a stanza from his sonnet ‘Recueillement’ [647]:

[647]  Pendant que des      mortels la   multitude vile

          while      that of-the  mortals the multitude vile

Sous  le  fouet  du      Plaisir,     ce   bourreau     sans      merci,

under the whip of-the pleasure, that executioner without mercy

Va     cueillir    des      remords dans la  fête      servile (1866)

goes to-gather of-the remorse  in     the festival servile

[647a] While the multitude of mortals under the whip of Pleasure, that merciless executioner, go to gather remorse in the servile festival

[647b] Now while the heedless throng makes haste to press

Where pleasure drives them, ruthless charioteer,

To pluck the fruits of sick remorse and fear (Alfred Douglas, 1919)

[647c] Now while the common multitude strips bare,

feels pleasure’s cat o’ nine tails on its back,

and fights off anguish at the great bazaar (Robert Lowell, 1961)

[647d] Now while the rotten herds of mankind,

Flogged by pleasure, that lyncher without touch,

Go picking remorse in their filthy holidays (Robert Bly, 1961)

[647e] While mortal multitudes with vile intents

By ruthless hangman Pleasure are whipped hence

To cull remorse in servile revelments (RdB, 1997)

[647b-d], all written by (reputed) English poets, join in differing starkly from the ‘prose style’ of [647a], provided as a ‘literal translation’ by an anthology editor (X.J. Kennedy, who is a reputed poet too). But they also diverge starkly in their approaches to style: Douglas seeking the most ornate style (e.g., ‘heedless throng’, ‘ruthless charioteer’), Lowell a lesser one (e.g., ‘common multitude’, ‘great bazaar’) and Bly a defiantly anti-poetic one (‘rotten herds’, ‘lyncher’, ‘filthy holidays’) and even abandoning meter for prose rhythm. Only my own version [647e] seeks to reconstruct a style parallel to Baudelaire’s, partly by using cognate lexical items like ‘mortal’, ‘multitudes’, ‘vile’, and ‘servile’ or French-seeming items like ‘revelments’, and partly by retaining his metaphors, which the three poets freely transformed. His ‘bourreau’ in [647], which suggests a state-sponsored executioner but equipped with a whip (‘fouet’) like a torturer, becomes a ‘charioteer’ [647b] (has a different profession but still uses a whip) or a ‘lyncher’ [647d] (executes with a rope and illegally), or just disappears [647c] (being the implied Agent using the ‘cat o’ nine tails’, a special whip). Such transformations can lead to incongruous collocations. If ‘fruit-pluckers’ were indeed ‘driven’ by a ‘charioteer’, they would presumably be riding inside and not running along under the whip in front. ‘Stripping bare’ would be done prior to a whipping, but is a peculiar Action at a ‘great bazaar’. A ‘lyncher’ would not hang a ‘herd’ of animals like cattle, who would already be dead anyway if they were ‘rotten’. Admittedly, we are now getting into the special problems of translation as well as comparative stylistics, and will take a closer look in VI.D.

16. If we examine interlingual discourse from the perspective of its participants, we encounter the issues of bilinguism or multilinguism, defined here not just as knowledge of more than one language but also as discourse practice in more than one. Although these phenomena are now more common worldwide than monolinguism is, most linguistic theoreticians, as Roy Harris and Alastair Pennycook note, still treat monolinguism as the norm. One type, called ‘coordinate bilinguism’, might have two systems stored fairly independently of each other, so that when one is activated the other is not. Another type, called ‘compound bilinguism’, might have the two systems stored in close correlation, so that both get activated together. Empirical evidence has been found for both types, raising the question of what conditions might favor the one over the other. The recently proposed °construction-integration model°, reviewed in III.250-56 and indicating that initial memory access is non-selective, makes the compound type more plausible. Also, this type would gain efficien cy by using shared entries (concepts, topics, schemas, etc.) in the total network of °knowledge about world and society° to support both languages; but we would need to account for the skillful sorting operations that keep bilingual speakers from continually manifesting interference, i.e., the activation of one system being impeded by control from another (cf. VI.21-30) and generating quasi-automatic lexicogrammatical mixtures, perhaps through the process of spreading activation cited in III.232 and III.246. I have observed in my own usage that I am most likely to switch languages unintentionally just after saying lexical items, especially long or complex ones, that are not clearly associated with the language I am speaking just then, such as a proper names (IV.23). Such anecdotal evidence suggests that going briefly outside one language system is a prime occasion for involuntary shifts to another system; the danger is much lower while staying inside one system.

17. We should also consider the respective design of the two languages a bilingual speaker knows. If the two designs differ strongly, more total data have to be stored but the risks are lower for interference, e.g., English and Chinese. If the two designs are quite similar, the reverse should hold: more compact data storage but also more interference, e.g., French and Spanish (after driving from Spain to France, I once groggily gave the silly time of day 2:30 as ‘deux heures et midi’, thinking of Spanish ‘dos y media’). These risks may be complicated by borrowings between otherwise dissimilar languages, witness Italian ‘capire’ versus German slang ‘kapieren’ (‘umderstand’), where you should be wary of saying  ‘kapieren Sie?’ rather than ‘verstehen Sie?’ to officials like Austrian police or customs officers at the border.

18. Evidently, similarities between languages in contact can raise efficien cy for text reception but lower it for text production. So an optimal design for bilinguism might be to have ‘compound’ reception and ‘co-ordinate’ production. But the routine interaction between reception and production should render this design difficult to achieve, at least without special training and intensive practice, as for simultaneous translating (cf. VI.21, 64, 74ff). My own competence certainly isn’t designed this way: after a week or two in Portugal or Brazil, my spoken Spanish degrades into the interlanguage known as ‘portunhol’.

19. Interlingual discourse can also promote the development of non-native language varieties, such as the varieties of English that have become entrenched in Africa and Asia where the ‘British Empire’ once was. After long being denied institutional recognition, some of these are finally being documented by the International Corpus of English directed until very recently by the late Sidney Greenbaum at University College London and slated to gather a million words of authentic texts in the local varieties of English in England, the U.S., Ireland, Canada, New Zealand, India, Singapore, Hong Kong, Philippines, Cameroon, Caribbean/Jamaica, Kenya/Tanzania/Zimbabwe, Nigeria, and South Africa. Such varieties are most obviously affected by local conditions in their lexicon, either by establishing otherwise unusual English collocations like ‘Himalayan blunder’ in Indian English, or by borrowing items from local languages like ‘otak udang’ (‘prawn-brain’ for a person acting stupid) taken from Malay into Singapore English (cf. VI.35; VII.344). The effects on prosody and grammar are more diffuse and subtle. Where British speakers might say an Exclamation with a prosody like

that shown in Fig. VI.1a, with a brisk rise and fall, I observed Singaporean students to say the one shown in Fig. VI.1b with the low opening syllable(s) leading into a sustained up-and-down curve toward the end of each tone group. The Singaporean stress patterns have probably been affected by the local languages: Malay, Tamil, and varieties of Chinese, chiefly Hokkien, Teochow, Cantonese, Hainanese, Hakka, and most recently Mandarin.

20. The grammar of Singapore English also reveals some interesting discourse signals. A prominent one is the multi-functional particle written ‘la’ or ‘lah’, which has been claimed to come from Chinese (compare the particle ‘le’ in Mandarin for signalling that ‘a state of affairs has special current relevance with respect to some particular situation’), and, among other things, is placed at the end of an utterance to signal of Certainty for commitment or emphasis, as in two samples I observed, [648] (carpenter) and [649] (student):

[648] I give it to you for nothing, lah

[649] I can’t help it, lah

The universal tag-question ‘is it?’ is used irrespective of the Person of the Verb Phrase upon which its depends, e.g. [650] (from a student-written musical) where a British speaker would prefer [650a]. Plausibly, this item was encouraged by invariant question particles in Chinese (‘ma’ in Mandarin) and Malay (‘kah’). A British speaker might mistakenly feel patronized, as if someone said ‘is that it?’ to suggest you have been devious or unclear.

[650] So we are feeling sorry now, is it? (from a student-authored musical)

[650a] So we are feeling sorry now, are we?

Although such features of local varieties have been loudly condemned by °language guardians with a colonialist ideology°  (cf. VI.50, 52, 60), they are far from disappearing in ordinary discourse.

VI.B. Direct evidence: Interference, borrowing, and code-switching

21. In the phenomenon of interference, mentioned in VI.16, the activation of one language system is observably impeded by control from another. This phenomenon contrasts with total bilinguism or multilinguism, wherein each system is successfully used on its own terms; probably, most people fall short of this ideal totality and encounter occasional interference, especially if the languages involved are rather similar (cf. VI.17f). Special training may also be helpful, especially for simultaneous translators (cf. VI.74ff).

22. If we view language as an abstract formal system (II.24, 47), interference appears to be mere noise; but if we view discourse as a concrete actual system , interference can offer helpful clues about discourse processing, e.g., about the °activation of concepts and expressions° (cf. III.248-55). Like speech errors, slips of the tongue, and miscues, interference might be a side-product of the routine efficiency gained from °self-organizing patterns of local activations°; the leading question would be how multilingual speakers can speak without interference (cf. VI.16). If one concept is closely connected to expressions in two languages, its °activation should briefly prime them both during text production°, and further operations could select the needed one and inhibit the other. If such operations run like those used for a single language to °construct and regulate active networks through parallel distributed processing°, suitable controls would need to inhibit the ‘wrong’ expressions from the non-selected language despite their meanings being relevant to the discourse topic (cf. III.246ff). A wrong item that is more closely connected to the concept in the speaker’s memory might win out over the right item and get passed along into production or at least get carried along and mixed with the right item, as when a German chap speaking English told us he had once been a ‘tellerdisher’ in a hotel (German ‘Tellerwäscher’ + English ‘dishwasher’).

23. An alternative account could be that activation does not work just with items, but with the whole subsector of the lexicogrammar for the current discourse domain, say, the set of items for the discourse of geometry in English. The discourse would be supported by some general but mild activation for the whole subsector in the selected language, but not for the corresponding subsector in the non-selected language. This mild activation could bias individual selections and reduce the competition between English versus non-English items. When an item occurs that is not clearly associated with the selected language, we might see the involuntary switching to a non-selected language (noted in VI.16) if, at that moment, the subsector bias is briefly deactivated.

24. Any such operations for inhibiting interference would be sensitive to the speaker’s fluency in the selected language. The etymology of this term would counsel using ‘fluency’ for a ‘flow’ of speaking with a suitable ‘pacing’ in the sense of III.204. Yet the term is often used instead for speech judged to manifest ‘correct’ grammar and pronunciation (cf. VII.324, 337). The two uses do not converge too well nor do they yield clear criteria. On the one side, pacing can reflect other factors besides your knowledge of the language, e.g., anxiety or concentrated attention in a dangerous situation such as an oral examination; and pacing can vary among different varieties of language, witness the smoothness and rapidity of Indian English , as contrasted with British English (VII.344). On the other side, ‘correctness’ is notoriously disputatious even among native speakers of English, simply because many choices among options are not specified by a frozen grammar but by ongoing emergent constraints (cf. VII.330).

25. The concept of fluency needs to be further specified for types of speakers and situations. For monolingual speakers, ‘fluency’ would designate a pacing that is not just smooth and rapid but appropriate  to the situation, e.g., slowing down for unfamiliar or difficult content or for an audience of non-specialists (cf. V.94ff). The standards would also be the appropriate  ones for the participants, and not just the ones prescribed by ‘grammar books’ or manuals of ‘elegant usage’. For multilingual speakers, fluency would designate a maintenance of appropriate  controls within each language; pacing would be balanced against standards, so that speech would be no faster than can be achieved without disruptive interference or degradation.

26. In this way, fluency could be realistically assessed relative to the speaker’s current knowledge of a language. It could be adjusted to accommodate the trade-off between speed versus accuracy (III.213.10) that is most acute in the early stages of non-native language instruction. It could also overcome the traditional tactic to uphold standards and suppress interference by drastically restricting the range of expression to a handful of prefabricated utterances (VII.324); the result is a seemingly fluent but wholly artificial ‘pseudo-discourse’, which tends to alienate the participants and to dilute their interest and motivation. The unconventional alternative would be to accept interference in early stages in order to build upon the native-language fluency as stipulated by a workable model of how discourse can strategically exploit control borrowed from a non-selected language. The model would need to specify the design for a series of provisional interlanguages to support real communication and to progress gradually but reliably from the native language to the foreign language (VII.336ff).

27. It would be vital here to distinguish whether interference actually disrupts communication or else is offset by °constraints from the participants’ knowledge of world and society°. Some interference may seem merely marked, e.g., florid [651] (native speaker of English), or quaint [652] (Russian). Moreover, the intended content is usually inferable from the situation, e.g., [653] (Arabic, notice in a Cairo hostel computer room) and [654] (Spanish, letter to me after a congress where my paper was read by the co-author). Interference is more damaging if comprehension would require some knowledge of the speaker’s native language, e.g., [655] (German) and [656-57] (Portuguese). The worst interference comes from translating ‘word for word’ with a bilingual dictionary, e.g., [658] (sign in a Spanish airplane) or [659] (notice in a hotel in Alicante or Alacant). English tourists can get the message of [659] by knowing the ‘payment’ such establishments prefer within the ‘politics of property and rent’ in modern Western cultures. Still, such extreme cases interference are rare in real discourse because even moderate fluency in a second language warns you not to rely on bit-for-bit matching alone. Instead, most of these cases are deliberately devised for marked or humorous  effects, e.g., [660] to cheer up my friend when we were stuck on a boat in windy weather and forced to go below deck, or [661] to vent my resentment on pompous, doltish German politicians like Helmut Kohl with their asinine English by caricaturing what they might say on a state visit to England.

[651] wir sind außer Speise geraten [= we ran out of food]

[652] I cannot tolerate your patience any longer [end of lecture]

[653] This lab is strictly allowed [= reserved] for the Hostel Residents. Please declare [= show] your ID card.

[654] I regretted you to be there [= I missed your being there]

[655] Because they are often ‘raunzig’ [i.e. grouchy], the Viennese have a bad call [= Ruf, i.e. reputation] in western Austria.

[656] finish water [= acabou agua, i.e., we’re out of water, not a command to finish off the water]

[657] close the light [= fecha a luz, i.e., turn off the light]

[658] in case of emergency pull the handle and discard [= arroje, i.e. push out ] the door

[659] In regime   from house rent of apartment,        the payment is for  anticipated.

   [= En regimen de    alquilar       de apartamentos, el  pago        es por adelantado,

    i.e., rent regulations  require that apartments be paid in advance]

[660]   self   me  pulls it   all too very

     [= selbst mir zieht es allzu   sehr, i.e., it’s too draughty even for me]

[661] I herewith vomit              this  stretch    to-the intercourse.

       [= hiermit  übergebe ich  diese Strecke dem  Verkehr,

             i.e.,  hereby   turn over  I      this   roadway to   traffic’

Admittedly, few interlingual jokes could top the grotesque comedy of nepotism, greed, and incompetence played out by the Kohl regime itself, where, as is typical of right- right-wing conservative governments, party loyalty and personal gain rank far above aptitude or intelligence (VII.47, 51).

28. The familiar cases of interference are the more local and formal ones affecting occasional Words or Phrases. Less familiar are the more global and functional cases affecting such factors as the organization of Processes and Participants or the distribution of focus. Viewed in a ‘functional sentence perspective’, English maps the pattern with the °Theme of content that is known, expected, or easily inferred, and the Rheme of content that is not° (IV.214ff) onto the patterns of ‘Subject + Predicate’, whereas Mandarin Chinese can follow Thematic priorities more closely. We notice some of the resulting interference when sample [662] was translated by a native speaker of English into [662a].

[662.1] Miss Xie is a graduate student from China. [662.2] Now she is studying American literature in the Department of English. [662.3] She likes American literature very much and has read a lot about the subject. [662.4] Some professors and students of the University know her.

[662a.1] Xié Xiàojié  shí  cong Zhōngguó lái       de    yánjiū-shēng. [662b.4] 

                       Xie Miss      be  from China        come mod  graduate student          3sng

              xián zái Yīngwén   Xì                 xué   Měiguó   wénxué. [662a.3] Tā    hěn  xǐhuān

                  now loc English   Department in study America literature              3sng very like

Měiguó    wénxué           guò   hěnduō  kèběn      guānyú zhè       kè.   

America  literature also read  perf  many    textbook  about    this   one subject

   [662a.4] Zái   dáxué       yǒuxiē  jiàoshòu  rènshi tā.

loc  university in some   professor know  sng

‘Xié Xiàojié’ is the Theme of [662.1] and is properly taken up by the co-referring Third-Person Singular (3SNG) Pro-Noun ‘tā’ at the start of the [662.2] and [662.3]. But the Phrase ‘guānyú zhé yī kè’ in [662.3], being Thematic, should go before the Verb and before the more Rhematic ‘hěnduō kèběn’. Interference from the prototypical English Subject-Predicate format was most glaring when [662.4] had the highly Thematic ‘tā’ after ‘yǒuxiē jiàoshòu’, Agents not mentioned before (though expected for a university); the student may have felt that putting the pronoun at the start would be as marked in Chinese as the English fronting ‘her some professors know’.

29. Conversely, this Chinese text [663] (from A Brief Introduction to Education in Shanxi) was translated by native Chinese speakers into the English text [663a]:

[663.1] Xīn    Zhōngguó chénglì yǐhóu Dǎng    zhèngfǔ       shífēn  zhòngshí 

New  China       found    after   party and government very    emphasize

jiáoyú       shīyé. [663.2] Jǐnguǎn zái   shí nián ‘Wénhuà Dà Gémìng’   zhōng  

education cause             though  loc  ten year  culture  big  revolution  in   

Shānxī    quán   guó       yīyàng, jiàoyù       shíyè  de      fāzhǎn

Shanxi and whole country same   education cause poss development

shòudào  yànzhòng  pòhuài.

suffer       severe      sabotage

[663a.1] Since the founding of New China, the Party and state have paid great attention to the development of education. [663a.2] Like other places in China, Shanxi suffered from the serious sabotage of the ten-year upheaval of the cultural revolution (1966-76) in education.

[663b.1] Since the founding of the People’s Republic of China, the Communist Party and the state have devoted great attention to the development of education. [663b.2] However, this development suffered serious sabotage in Shanxi, as elsewhere in China, during the ten-year upheaval of the Great Cultural Revolution (1966-76).

In [663a], some cultural discrepancies are conspicuous, e.g., Westerners might need to be informed that ‘New China’ and ‘the Party’ are ‘the People’s Republic of China’ and ‘the Communist Party’. Less conspicuous is the uncertain handling of Theme and Rheme. The opening in [663a.1] takes ‘the founding of New China’ for the background and puts ‘great attention to education’ in the foreground. The first stretch of [663a.2] leads the reader to expect a confirmation, e.g., ‘Like other places in China, this attention promoted education in Shanxi’. But a contradiction is stated instead in ‘serious sabotage’, while ‘education’ gets pushed along to the very end of the sentence, as if ‘education’ had been the domain where the ‘cultural revolution’ took place. I would propose to use the unmarked English sequence whereby the Rheme of one Sentence becomes the Theme of the next, hence the recurrence ‘development’ - ‘development’ [663b.1-2], making it unnecessary to repeat ‘education’; end-weight focus can go to the Rhematic ‘sabotage’ and its cause. A junctive like ‘However’ can set up the contrast of [663b.2] against what [663b.1] leads you to expect.

30. Instead of just deploring or punishing interference, we should explore its sources as a natural side-product of the capacity of discourse processing to °integrate multiple constraints without explicit controls°. The constraints of a non-selected language in which the speaker is also fluent can easily intrude. We might obtain some valuable evidence both about the gradual evolution of discourse skills while learning a second language and about the rapid evolution of current versions of language during discourse (cf. III.128-36). For example, interference might prove to fluctuate according to ongoing resource demands, increasing under competition from stress, noise, and fast timing, or decreasing with support from motor precision, attention, or feedback. If so, we can design models for language learning under conditions that help learners to control interference without severe anxiety and to steadily phase it out later on (cf. III.336f).

31. Unlike interference, borrowing deliberately appropriates specific items from another language and frames them into its own organization during communication. In nonce borrowing, an individual text producer makes an isolated appropriation from the °framed language°, usually because a comparable term in the °framing language° either is not available and needs explaining [664] (native speaker of Lithuanian) or else would seem marked [665] (German, where Noun and Verb do not coincide as in English ‘mail’).

[664] Also worth a try are gira (a refreshing beverage made by fermenting grains of fruit […] But beware of samagonas, a powerful and often nasty-tasting home brew of cheap wine!

[665] wir haben dir schon etwas ge-e-mail-t [= we already sent you some e-mail]

32. Communal borrowings are accepted by a whole community. These are frequent among immigrants who are not schooled in foreign languages and retain their native language in their new environment. The popular play Bordynkari (from English ‘boarders’), adapted by Jaroslav Psenka (1908) from a novel of the same name by František Škaloud (1898), portrays a Czech community in Chicago using English words they have largely adapted to their native sounds, spelling, and grammar, e.g., ‘bajbaj’ from ‘bye-bye’, ‘pedlak’ from ‘peddler’, and ‘departmentní obchod’ from ‘department store’. Czech supplied the framework of the utterances, e.g. in [666] and [667], the latter with the Sentence Subject getting focused end weight in contrast to the English version.

[666] jak to bude s tím  čenčem za ten dolar? [what about the change for my dollar?]

[667] pracovali tam plumbari a papírovaši [plumbers and paperhangers worked there]

When borrowings are selected for prestige , the original pronunciation and orthography are most likely to be preserved, as in the German uses of English ‘down’ (for mood), ‘high’ (for altered awareness), and ‘easy’ (for feeling at ease). But prolonged ordinary use favors assimilation to the framing language, e.g., English ‘orangutan’ (pronounced to rhyme with ‘I rang and rang’) from Malay ‘orang hutan’ (‘person [of] forest’), or Malay ‘bas sekolah’ from English ‘school bus’. Eventually, such items are no longer regarded as borrowings at all and are barely noticed in discourse; certainly nobody considers Malaysia’s national ‘Parlimen’ undignified because its name is not spelled or pronounced like English ‘parliament’.

33. Code-switching is a far more extensive and less deliberate interface, where speakers (and much less often, writers) episodically switch between the framing and the framed language. Even when the proportion of the framed language is quite high, the framing language, usually the speaker’s more fluent one, is recognized by providing the patterns. In general, it provides the units matching the morphemes and syntagmemes, and the Function Words for grammaticalizing, while the framed language is limited to episodic lexicalizing of Content Words, especially ones not used in the discourse as Adjuncts to expressions in the framing language. These tendencies endow code-switching with some regular emergent constraints despite its frequently improvised quality. We can see some evidence in recordings of Chinese students in the U.S. collected by Longxing Wei, where the observed data could not be changed as shown: the English items are Cores with Adjuncts (e.g. ‘my advisor’) rather than Adjuncts of Chinese Cores (e.g. ‘*my dǎoshī’) [667], and the groups are kept together rather than interrupted [668]. More deliberate nonce-like switching can be done to achieve marked affects, such as the striking or humorous graffiti popular in German-speaking cities like Vienna [669-70].

[668] qù jiàn my advisor  tā zuìjìn very busy  vs. *qù jiàn my dǎoshī  tā zuìjìn very máng [go     see my advisor, lately he’s very busy]

[669] business building qián yǒu parking lot vs. *qián business building yǒu parking lot [in front of business building there’s a parking lot]

[670] No Atomstrom in my Wohnhome [no nuclear power in my residence hall <= German ‘Wohnheim’]

[671] Mehr power to the Bauer, mehr Recht for the Knecht [more power for farmers, more rights for laborers]

34. As suggested by the Czech and Chinese examples, the leading motive for code-switching is a multilingual setting, e.g., in displaced cultures of recent or ethnically distinctive immigrants or in former colonies like the Philippines (cf. VI.84; VIII.33ff). A less pervasive but still significant motive would be educational and professional institutions in which certain technical subjects are taught or discussed in non-native languages such as English. There, many switched-to items are technical terms, e.g. geology in [672] and medicine [673-74], recorded by Ibrahim Khidhir Sallo at the University of Mosul, Iraq. For [673], he suggested that the lecturer might also be switching to English to ‘avoid socially tabooed’ expressions, whereas the speaker of [674] was merely trying to appear fashionable. Back in 1977, a decree was issued mandating Arabic as the medium of instruction at the university level in Iraq; and since then, Britain, the old colonial master, and the U.S. have become bitter military enemies and imposers of painful embargoes. Yet data like Sallo’s show how popular the use of English still is  (? now has to serve for hamza, the unvoiced glottal stop).

[672] a -sukhuur imma ?an takuun rusuubiyya sedimentary, ?aw muta awwila

    metamorphic, ?aw naariyya ?ay igneous  [the rocks are either ...or...or...]

[673] bi aala wa da lamma nsawwii vagina lavage nshuuf aakuu blood bil-fluid

 illii raa nis aba min al-vagina, fabhaay al- aala mumkin nshuk innahaa pregnant 

[when we carry out vaginal lavage, we see whether there is blood in the

 fluid which is taken from the vagina; in that case, we can suspect that she is pregnant]

[674] naakhudh ?al-inaath in the early morning [we take the females...]

Even where code-switching is widespread, governments and institutions have seldom recognized and exploited its potential for an integrative program in multilingual education, partly because modern education began with monolingual methods and partly because languages typically differ sharply in prestige (cf. VII.29, 57). Such programs are urgently needed in a world where multilinguism has surpassed monolinguism (V.97; VI.16; VII.12329).

35. As for many pairs of categories in language or discourse, borrowing and code-switching do not form a neat dichotomy. Singapore English has many widely-shared borrowings from the Chinese and Malay whose usage is not felt to constitute code-switching, e.g., ‘ang pow’ for the gifts distributed among the children of visitors at Chinese New Year. And most borrowings are not fully Anglicized but retain the original pronunciation, e.g. ‘otak udang’ (‘prawn-brain’) cited in VI.19 being ['otak 'u:dang] and not, say, ['outæ:k 'ju:dæ:ng]. The influence of Bahasa spoken in Peninsular Malaysia, wherein final ‘-k’ is now realized as a glottal stop, might yield the pronunciation ['ota?].

36. To a discourse participant unaccustomed to them, interference, borrowing, and code-switching may seem to disturb communication. However, they do not do so in the cultural contexts where we mainly find them. [675] is a written note delivered to a Manila householder by her gardener’s wife, a native speaker of Tagalog-based Filipino. For elitist English-educated Filipinos, this would be dismissed as ‘Bamboo English’. But the gardener’s message is hardly less clear than if it had used the ‘standard Filipino English’, the term preferred by Teodoro Llamzon, e.g. [675a].

[675] Please excused my husban that caused he is not to worked last Friday. Because he had sicked of two days.

[675a] Please excuse my husband because he didn’t work last Friday, because he was sick for two days.

The formal details, e.g., whether ‘he is not to worked’ should correspond to ‘he did not work’ or ‘he did not come to work’ would hardly matter to the gardener and the householder.

37. As we often find, a text that seems objectionable on formal grounds is quite functional for the context of situation and the agenda (cf. II.75; V.50). In the °lexicogrammatical economy°, an Emotive Cognition (‘excuse’, i.e., regard without annoyance) is plainly requested for a Negative Enactive (‘not coming to work’) caused by a Pejorative Development (‘getting sick’); the only doubtful point is the Temporal Circumstance, which the text places into the past by more than one day (hence not ‘yesterday’) but less than one week (hence ‘last Friday’). In fact, the top of the paper bore the date ‘July 8/89’, which was a Saturday, but the ‘8’ for the day had been visibly written over a ‘9’. We can surmise that the message was written Sunday the 9th and then backdated to save face by seeming more punctual and assiduous; the writer failed to notice that the revised date would call for ‘yesterday’. This conclusion is plausible also because ‘two days’ could justify not coming on Saturday in exchange for Friday.

38. Regarding intertextuality, [675] follows the well-defined text type ‘excuse note’ dating from school days and requiring that the sick person not be the writer (hence not ‘Please excuse me’). For children, this requirement reflects the social and legal authority of parents to judge between health versus illness of offspring. An adult gardener should have his own authority, but ours declined to use it even though he did know how to write, witness his code-switching note [676] delivered two days later and dated July 11. [676] follows a less well-defined text type we might call ‘maintenance expense note’. The handwriting was different and less legible than in the previous note, and the code-switching indicated a lower fluency in English than the wife’s, limited mainly to stock collocations of politeness (e.g., ‘thank you very much’).

[676] Pahasa   longmower  P          80.    Thank you very much.

        sharpen   lawnmower  pesos eighty [i.e., it cost eighty pesos to sharpen  the

         lawnmower]

The two text types suited two different agendas for the gardener’s professional role. When he was not fulfilling it in being sick and absent, he addressed his employer through a third party in a more official note, neatly written all in English and backdated. But when he was fulfilling it in maintaining the equipment (‘longmower’ as folk etymology: it mows long grass), he felt socially competent to address her himself in a casually written note with code-switching. The householder duly accepted the messages and was satisfied by the display of concern.

39. To estimate how interference from Filipino may have exerted control, we can compare the original [675] with a plausible Filipino version [675b] (aux = Auxiliary; det = Determiner).

[675b] Ipagpaumanhin          ang aking asawa   sapagkat  siya’y   hindi

        be-pleased-to-excuse det my    spouse  because   he-aux  not  

nag-trabaho noong Biyernes, sapagkat siya ay nag-kasakit   ng datawang araw.

past-work   on        Friday      because  he-aux past-be sick for  two          days

For instance, since the preposition ‘ng’ corresponds to both ‘for’ and ‘of’ in English, the phrasing ‘of two days’ rather than ‘for two days’ might seem plausible. Or, Filipino Verb forms need endings, and the English ones in [675] were outfitted with the common (though inappropriate) Verb ending ‘-(e)d’. The particle ‘ay’, contracted to ‘-y’ after a form ending in ‘-a’ (compare English ‘he’s’), is an Auxiliary (aux) functioning like both ‘is’ and ‘does’ in English Verb Phrases; ‘is not work’ might get used for ‘did not work’. In sharp contrast to English, the Non-Finite Process Verb in Filipino is marked for Tense (here with the Prefix ‘nag-’) while the Auxiliary is not; hence ‘is not to worked’ might appear normal. Also, the Predecessive ‘had sicked’ would indicate that the ‘sicking’ occurred before the ‘not to working’; and since ‘nagkasakit’ is a Process Verb, not ‘be’ + State, a conjugated form like ‘sicked’ makes sense.