John McWhorter responds

« previous post | next post »

Some clarifications about my Wall Street Journal article, which seems to have led to some misunderstandings among Language Log’s readers (as well as over at Languagehat). Since the readers here are the most well-informed audience that piece will ever reach outside of professional linguists, I thought it’d be useful to clarify what I based the observations in that piece on.

Contact and simplification

It is unfortunate if I seemed to be making the rather absurd point that language contact guarantees simplification. I could not be someone who thinks of Sally Thomason’s monograph on language contact with Terence Kaufman as a Bible and harbor so messy a notion – why, after all, aren’t all languages pidgins by now? Most language contact results simply in mixture of various kinds.

However, I refer to a particular kind of language contact: where acquisition by adults is so extensive that their non-native rendition of the language is transmitted to new generations and becomes the norm. Peter Trudgill estimates the proportion of adult learners necessary to this process as roughly 50%. There has been a great deal of work distinguishing this kind of contact since the Thomason & Kaufman volume, by, for example, Kusters (2003), myself (my book Language Interrupted, consisting of several detailed case studies) and Trudgill (2011). I would venture to say that the empirical validity of this analyis of certain cases of language contact is largely unarguable; in 15 years there have been no conclusive smackdowns of the basic mechanism proposed by Kusters, Trudgill, and well, that other person.

The question would be: if a critical mass of adults learn a language and there is no educational system or media to maintain the dominance of the older form, why would the language passed on to new generations not be shorn somewhat of its more learner-unfriendly traits? It’s indisputable that this is what creates a pidgin (and/or creole). That there would not be intermediate cases is what we would not expect. English, Persian, Mandarin are exactly such cases.

Caveat: the idea is not that languages like these were “creolized.” The proliferation of languages as radically simplified as creoles was mainly at one stage in global history, with massive and rapid population movements of adults ending up in the bizarre situation of then speaking a language they knew only partially more to one another than to their rulers. That was a circumstance occasioned largely by plantation slavery and similar situations, and was a recipe for severely abbreviated acquisition, followed perhaps by the reconstitution of its result into a new language. In shorthand, this is the famous pidgin-to-creole scenario.

However, the filtering that created English, Persian, etc. was hardly as extreme as creolization, nor was the “third wave” of simplification I refer to today with immigrant varieties. In both cases, there was more what one might call “streamlining.” Linguistics lacks a handy term for simplification of intermediate degree. This creates endless confusion, such as whole discussions as to whether English was creole, whether modern colloquial Arabics have their roots in a “pidgin” Arabic and so on. Francophone linguists’ term “vehicularization” ought be more popular.

English and simplification

Truthfully, however, nailing down that ever elusive general definition of grammatical complexity is not necessary to the basic point about what happened to, say, English (although I attempt such a definition in Language Interrupted, a book read by essentially no one due to my own laxity in publicizing it). Example: English lost not only affixes but much else: it lost more of the Proto-Germanic inheritance in myriad ways than any other Germanic language. My example from the trade book Our Magnificent Bastard Tongue is likely useful here: English has Did she say to my daughter that my father has come alone and is feeling better? But German has “Said she to my daughter that my father alone come is and himself better feels?” and it comes out roughly that way in the other Germanic languages. Loss of affixes is only the beginning of what distinguishes those two sentences — and English’s taking on do-support doesn’t begin to equal the massive volume of loss of Proto-Germanic material. Something like this begs for explanation.

The most plausible one is non-native acquisition by Scandinavian Vikings, as increasing numbers of specialists are now stating (D. Gary Miller is especially useful in his latest). Note: the French were not the culprits – there were never enough of them in England to affect speech among the Anglophone masses, and they switched to English rather quickly anyway. Thomason & Kaufman, for example, was authoritative on this. Mainland Scandinavian languages later became a close second to English in this regard – but only that; English losses were much greater — because of Low German speakers’ non-native acquisition; Trudgill makes this case most accessibly.

What About Russian?

A crucial caveat, though: this kind of acquisition was most impactful before widespread education and literacy. Russian has been used as a second-language quite a bit without being simplified, indeed – but its spread has been reinforced to a large extent by formal education, literacy, and then media. Certainly there have been non-native varieties of Russian spoken in a great many places – but they almost never reach print and will never become the standard. “Broken” English took over in a country where most people were essentially illiterate, there was barely a such thing as school, and then after a long period when only French was written and the old tradition of writing in a high West Saxon Old English became a mere memory, it felt natural to start using “on the ground” English on the page.

Today, it is much harder for non-prescriptive varieties to be reinterpreted as prestigious ones in this way. The “immigrant” Swedish now spoken by children of immigrants will never oust standard Swedish or affect it in any real way, whereas the “immigrant” Norwegian spoken by Low Germans several centuries ago became the Norwegian norm in the area (whereas Scandinavian dialects further north such as the unfortunately obscure Elfdalian retain Old Norse’s three genders, etc.). We moderns perhaps have to strain a bit to imagine worlds where language was primarily oral and our prescriptivist sense of language barely existed.

Mandarin Worldwide

I can hardly claim to base my speculation that the Chinese will rule the world in English on any unassailable powers of prediction. However, I do think that current conditions will powerfully affect the future – to wit, English happens to have settled in at a time when communications and media could make a language more prevalent (and thus often perceived, for better or worse and often the latter, as “cooler,” also) than ever before in human history. The serial language takeovers artfully described in, for example, Nicholas Ostler’s books cannot, I think, be taken as evidence that English must be ripe for replacement. Path dependence – i.e. the QWERTY problem or even English spelling – is powerful stuff.

Moreover, instant translation devices will affect things, but not the sheer fact that English will retain its visceral associations – with the cosmopolitan, with money, even glamour. I say that not as some kind of booster, despite that English happens to be my language – it just strikes me as likely. Now, to the extent that Mandarin will be learned more by others in the future, which I do not deny, I think that the writing system as it stands now would preserve that as an elite activity cherished more by language fans and people especially fascinated by China than the general public. However, as one poster notes, pinyin makes Chinese more accessible, and if the Chinese switched to that, then …

…well, nothing much different, is my guess, because then, one still has to process the tones. Let’s recall that it isn’t that each Mandarin syllable can have only four meanings each corresponding to one of the four tones – on each tone the typical word has plenty of homonyms. Take a look at what, say, yi can mean – with countless meanings on each tone — to get a sense of it. In that light, it is useful to note that adult learners have a notoriously hard time with tones: even speakers of closely related tonal languages, when they come together and create a lingua franca, often shed using the tones to distinguish words and grammar by themselves – there are examples in Africa, for instance. To wit: I hypothesize, based on how such matters have come out in cases worldwide, that if Thai plantations had been worked by Chinese and Hmong slaves, the resultant creole would still not have used tones to distinguish words or encode grammar.

Now, one thing that could happen is that a worldwide Mandarin could emerge that relied on compounds even more than the modern language does (which is massively). That is, instead of bi meaning “nose,” “pen,” “than,” “certainly,” “currency,” “finish,” “avoid,” etc. depending on the tone, maybe a World Mandarin would not use tones, but bi alone would be reserved for “nose,” and “pen” would be “writing bi,” “than” would be “compare bi,” “finish” would be “end bi” and so on. This is how the actual language already works to an extent – a more universally approachable Mandarin might make it the rule.

Who knows? But then, today’s standard Mandarin would stand in judgment, unlike it would have millennia ago in a different technological world. That toneless Mandarin stuffed with novel compounds would sound horrific to most natives – how would that change? Perhaps I lack imagination ….



21 Comments

  1. Pat Barrett said,

    January 29, 2015 @ 1:45 pm

    The books mentioned in John McWhorter's response will clarify a lot, but right now I'm reading his The Missing Spanish Creoles and Defining Creole, and what strikes me is the way he deftly threads through every possible "yes, but…." capable of being thrown up and either dismissing it with evidence or showing how it can be taken into account. Some of the comments here and at Language Hat seem to stem from a fuzzy notion of what makes a language or language feature complex or simple or reduced, I.e. we have no definitions i common. In addition, the external linguistics matter a great deal, i.e. the setting, and the complexities of the Atlantic Slave Trade are dealt with decisively in McWhorter's books on Creoles arising out of that trade. Reading these books would answer a lot of the questions and objections raised.

  2. S Frankel said,

    January 29, 2015 @ 2:19 pm

    I haven't read D Gary Miller's book, but the notion of Scandinavian influence being responsible for the simplification of English is strange, because Old English and Old Norse (or whatever you want to call the latter) were quite similar, at least in their modern normative forms. The biggest difference would be in the definite article (Scandinavian had, and still has, both pre- and postposed forms.

    Furthermore, the last part of this is dead wrong:

    German has “Said she to my daughter that my father alone come is and himself better feels?” and it comes out roughly that way in the other Germanic languages.

    That order is peculiar to modern German only. Older German didn't kick the verbs to the ends of clauses; neither do the other modern Germanic languages, including Yiddish. The only difference between English and the other Germanic languages in that example is the form of the English question, which some people feel is due to Celtic influence. (Welsh has the identical construction, but it pops up in English too recently to be an easily assumed Celtic substratal influence.)

    (Oh, one minor difference, the word for "feel" here would be reflexive in at least some other Germanic languages.)

  3. S Frankel said,

    January 29, 2015 @ 2:23 pm

    Sorry – to clarify (and to click the "notify me of follow-up comments" box which I forgot to do before): All the Germanic languages except English have V2 order. German and Dutch (and not all the local varieties of the former; I don't know about the latter) kick some of the verbs to the ends of some clauses (subordinate clauses, mostly), but that is an innovation.

  4. J. W. Brewer said,

    January 29, 2015 @ 2:52 pm

    I very much appreciate Prof. McWhorter's willingness to engage further on these issues in this forum. I have I guess two outstanding primary concerns.

    1. It is not clear to me that the "critical mass" etc. conditions for the streamlining/vehicularization process outlined above are likely to obtain over the course of the next century for many/most currently-strong languages. This latest post seems to concede that for Swedish (and, as matter of recent history rather than future expectation, for Russian). So I'm left puzzled as to the basis for the prediction that the world is going to have fewer but simpler languages. Which languages are likely to experience the proper conditions (in terms of a truly massive influx of new adult learners swamping the stabilizing effects of formal education etc etc and thus influencing the "mainstream" version of their newly-acquired language going forward) for further streamlining?

    2. If we set aside potentially unproductive terminological quibbling about how to define and measure "complexity," is the point of the comparison between the English sentence ("Did she say" etc.) and its German equivalent that the English pattern is objectively more "streamlined" for some definition of "streamlined" that is capable of generating testable hypotheses? For example. do immigrants from a range of different L1 backgrounds not particularly similar to either English or German find it easier ceteris paribus to master English syntax than German syntax? Is there any good empirical research (I can see a lot of methodological problems in trying to adequately control for the many other variables, but that doesn't mean smart scholars haven't tried to solve them) on questions like that?

  5. Jim said,

    January 29, 2015 @ 5:39 pm

    "…well, nothing much different, is my guess, because then, one still has to process the tones."

    Well with English there are all the vowels to deal with, often with comical results. I mean, that beach is hard.

    "That there would not be intermediate cases is what we would not expect. English, Persian, Mandarin are exactly such cases."

    I wonder if there is a subsets of Sprachbunds where one language is the center and the others orbit it, whereas in other types of Sprachbund the member languages are on an equal enough basis that they freely swap equally.

  6. Matt said,

    January 29, 2015 @ 7:15 pm

    I mean, that beach is hard.

    Most people do manage to wreck it, though, if it's nice enough.

  7. Y said,

    January 29, 2015 @ 7:53 pm

    I don't see why people shifting from one tonal language to another would lose tone, as in the case you imagine with Thai, Hmong, and Chinese.

  8. JS said,

    January 29, 2015 @ 8:03 pm

    Of Mandarin, "on each tone the typical word has plenty of homonyms"
    The typical Chinese word is more than one syllable; the author is explicitly aware of this but still speaks of a non-existent homophone problem in Mandarin. Or should the word "word" above be "syllable/morpheme" instead, creating a true-ish statement that would instead be a comment on the relatively simple phonological system and syllable structure of Mandarin and other languages of the area? If so, would that make such languages complicated (different words might arguably sound more similar to each other as the same sounds and syllables are reused at relatively high frequencies), or simple (fewer sounds = easier to learn)?

    "speakers of closely related tonal languages, when they come together and create a lingua franca, often shed using the tones to distinguish words and grammar by themselves"

    I am interested in the examples in Africa to which the author refers. In China and Southeast Asia, far from being subject to reduction or elimination, tone is *the* single feature most illustrative of the power and pervasiveness of language change by contact. In that linguistic context, the idea of two languages featuring lexical tone yielding a toneless hybrid seems an utterly bizarre one.

    More generally, I think the folks emphasizing the inherent difficulty of Mandarin, etc., need only to see a young non-native speaker acquire it effortlessly in an immersion situation. And I don't know why the emphasis on pronunciation, rather trivial in my mind compared to issues of syntax-semantics interface, where Chinese is a paragon of virtue next to English and most of the rest.

  9. Alicia said,

    January 29, 2015 @ 10:13 pm

    @ Y:
    He says he makes the prediction of the loss of tones based on what happened when other tonal languages have creolized together.

    @ John McW:
    I read Language Interrupted, so there! And you plugged in in every foot note in… One of your more accessible books, which I read and then ILLed both this book and the "Missing Spanish Creoles" and slogged my way through them. So you can't say you didn't try.
    Oh, and I xeroxed your stuff about "Me and Billy" and shoved it up the noses of all the "Billy and I" snobs I know, and have benefited from many of your other erudite anti-perscriptivist analogies, so much thanks and write some more books please.
    PS: I bought "Marvelous Bastard Tongue" for my 12 year old niece, then on reviewing it realized that it assumed the reader already knew the conventional story line (had heard of the Norman invasion etc etc), so I got her Bill Bryson's Mother Tongue to fill that in, since I remembered having found it accessible and not much more inaccurate than most other books. After doing presents she wandered off do homework and I entertained myself rereading the Bryson. Turns out it has a tremendously "educational" (I mean TREMENDOUSLY educational) chapter on cuss words. Oh well. Hopefully they'll both get lost under her bed until she's 14…

  10. Apollo Wu said,

    January 29, 2015 @ 10:33 pm

    The emergence of tone in ancient China probably linked to the desire of extending the use of monosyllabic approach. Such effort only delay the widespread use of polysyllabic words, as it would be unwieldy to keep increasing the number of tones. Mandarin has already fewer tones than Cantonese, and there seems to be a process of switching towards the first tone as in that for 微 and 期,(as in 微软,假期) both have changed from second tone to first tone. Language change is a slow process and the trend appears to be in favor of giving up the use of tones for specific meanings. Today, one hardly rely on tonal information when listening to fast speaking broadcasters.

  11. Mark Mandel said,

    January 29, 2015 @ 11:05 pm

    «instead of bi meaning “nose,” “pen,” … etc. depending on the tone, maybe a World Mandarin would not use tones, but bi alone would be reserved for “nose,” and “pen” would be “writing bi,” …»

    This immediately reminded me of my first encounter with pre-nasal raising in the US South, which makes homophones of "pin" and "pen". I clearly remember my shock (I was 16 or 17) at hearing them routinely distinguished as "a ink-pen" and "a stick-pin".

  12. Victor Mair said,

    January 30, 2015 @ 1:22 am

    The tones are not immutable.

    This is an issue that has frequently come up on Language Log:

    "When intonation overrides tone"

    http://languagelog.ldc.upenn.edu/nll/?p=4673

    "Archive for Tones"

    http://languagelog.ldc.upenn.edu/nll/?cat=215

    Just last week I was called out for following the old, traditional second tone pronunciation of a character that had shifted to first tone on the Mainland:

    http://languagelog.ldc.upenn.edu/nll/?p=17258#comment-1489429

    Apollo Wu is right to point out that a lot of second tone morphemes in Taiwan Mandarin have been shifting to first tone in Mainland Mandarin. This is a phenomenon that I had noticed long ago. I have my own pet theory about why it is happening, but I've never heard anyone give a convincing explanation that is widely accepted for this change of other tones to first tone.

    Apollo further says: "Language change is a slow process and the trend appears to be in favor of giving up the use of tones for specific meanings. Today, one hardly rely on tonal information when listening to fast speaking broadcasters."

    That's a very interesting, and I think valid, statement. Serendipitously, it also matches something that came up today in a discussion that was prompted by Randy Alexander, who asked this simple question: Have you ever heard bēibǐ 卑鄙 ("despicable; mean; base; contemptible") pronounced as bēibì? That would be a shift from third to fourth tone on the second syllable.

    Here are some comments from several of my grad students from the PRC:

    =====

    Yes, especially in the phrase bēibǐ xiàliú [–> bēibì xiàliú] 卑鄙下流 ("nasty; vile; repulsive; lousy; dirty low-down"). I guess it's easier than pronouncing it in the third tone.

    Many Chinese native speakers pronounced bǐ 鄙 as bi4. Very common.

    Yes, I sometimes heard people pronounced it as bei1bi4. I guess it is hard to tell the third or fourth tone when people read them fast.

    =====

    These remarks were made by highly educated, intelligent native speakers of Mainland Mandarin. It's interesting that the first respondent explains the change from bǐ to bì in terms of ease of pronunciation: it's easier to pronounce a fourth tone than a third tone. This matches well something that my wife, who was a talented Mandarin teacher for more than half a century, used to say about the third tone. She said that to pronounce it properly requires a real, physical effort to reach down low in your vocal register, especially when you do a full third tone with a slide downward and then rise upward. She said that most of the time people just don't do a full third tone because it takes too much effort and time. Instead they do what she called a "half-third tone", where the speaker just starts down near the bottom of his / her register without worrying about the dip. Incidentally, I think that the physical effort needed to stretch down to the bottom of one's register on a succession of third tones is what triggers the tone sandhi where the preceding of two third tones switches to second tone, which is similar to a truncated third tone (i.e., just the latter, rising part).

    What the third respondent says about tones and speed of speech matches Apollo's other point perfectly. Tones consist of a changing pitch contour, and that takes a certain amount of time to produce. (The exception, of course, is first tone, where there is no change in pitch contour.) So, when one is speaking very fast, it becomes virtually impossible to fully enunciate the tones; they get short shrift. Apollo and the third respondent are of one mind about this. To enunciate the tones fully and clearly necessitates a fairly slow and measured pattern of speech.

    The above adumbrates my views about why tones are shifting, especially toward the first tone.

  13. christoll said,

    January 30, 2015 @ 1:31 am

    The only difference between English and the other Germanic languages in that example is the form of the English question, which some people feel is due to Celtic influence. (Welsh has the identical construction, but it pops up in English too recently to be an easily assumed Celtic substratal influence.)

    Too recently? Only if you assume that the Old English written language – as written by the elite of Anglo-Saxon society – was more or less the same as the spoken language of the vast bulk of the people, who didn't write anything at all.

    It's actually quite likely that the spoken language of most people in England was markedly different from Old English as it appears in the documentary record, even before the Viking conquest – just as everyday spoken French today is quite different from formal written French, for example. But the deviant features of the spoken language would not have appeared in written English until the Anglo-Saxon elite had been ousted and a new elite had risen to take their place. Those Celtic aspects could have been there lurking in the spoken language of the peasant masses the whole time.

  14. Bert Remijsen said,

    January 30, 2015 @ 5:53 am

    This is a reaction to the following statement: "I hypothesize, based on how such matters have come out in cases worldwide, that if Thai plantations had been worked by Chinese and Hmong slaves, the resultant creole would still not have used tones to distinguish words or encode grammar." Surely, contact between tone languages is prone to yield another tone language. I cannot think of that particular socio-economic context. But consider the Austronesian languages, which on the whole do not present lexical specification of tone. The handful of Austronesian tone languages are to be found primarily in regions where the Austronesian language family borders languages that present tone: Southeast Asia and New Guinea. It is contact, rather than internal development, that has been the main source of tonogenesis in this language family.

  15. GeorgeW said,

    January 30, 2015 @ 6:37 am

    y: "I don't see why people shifting from one tonal language to another would lose tone . . ."

    I could be related to the emergence of the unmarked feature in SLA. As an example, studies have shown that Farsi learners of English devoice word-final obstruents (an unmarked feature) even though both languages have word-final contrasts.

  16. DMT said,

    January 30, 2015 @ 6:53 am

    I would be curious to hear from some of the Siniticists here about the current state of opinion on McWhorter's description of Mandarin as a creole. (From Language Interrupted: "More precisely, the data suggest that the southern Chinese varieties were impacted by extensive language shift amid proficient bilingualism, while Mandarin resulted from robust but incomplete acquisition.")

  17. S Frankel said,

    January 30, 2015 @ 10:03 am

    @ Christoll – Right. The only issue is that it's an assumption. There's no documentary evidence for a Celtic origin for the emergence of the English use of "do" as an auxiliary. It's a neat explanation, but it's not an open-and-shut case.

    Obviously, fashions change in historical linguistics, as in everything else. It used to be assumed that OF COURSE the auxiliary "do" had nothing to do with Celtic, mostly on the grounds that there are incredibly few Welsh (or "British") loanwords in English except for place names.

    And "do" can be used as a verb of low semantic content in other Germanic languages, typically in answering yes-or-no questions. German: Kommst du morgen? Ja, das tu' ich. ( lit: Comest thou tomorrow? Yes, that do I.) This construction is a bit perpheral in German, but obligatory in most modern Scandinavian, and there are hints of it in Old English. So, we have a widespread use of an auxiliary "do" in Germanic that can't be attributed to Celtic.

    Incidentally, in Breton, we also have a Celtic auxiliary "do" (as a periphrastic) that can't be attributed to English.

    Beyond that, it's just a bunch of guesswork. One generation's reasonable assumption is another's obvious mistake.

  18. Y said,

    January 30, 2015 @ 4:57 pm

    'Marked' features persist into creoles and other contact varieties, when all languages involved have them. French-lexified African and Vietnamese creoles have nasalized vowels; Dravidian-Indic contact languages have retroflex stops. I don't see why tone would be different. Certainly in borrowing situations from one tonal language into another tones are not stripped away; see Martha Ratliff, Timing Tonogenesis: Evidence from Borrowing (BLS 28S, 29, 2002).

  19. E Gilman said,

    January 31, 2015 @ 6:10 pm

    I am curious as to where Middle English got it SVO word order. I don't think the Vikings could have done that, could they?

  20. S Frankel said,

    January 31, 2015 @ 9:45 pm

    Right, all the Scandinavian languages are, and have always been V2. One more data point: Old French is also usually reckoned to be V2, due to Germanic influence. Don't know how artificial that was, though; the language is a bit of a literary construct, and its relation to actual speech isn't always clear. I don't know anything about Norman French.

  21. Sybil said,

    February 1, 2015 @ 1:28 am

    McWhorter's book "Language Interrupted" is available online for those of us with a valid NYPL card. Maybe other libraries as well?

RSS feed for comments on this post