THE STRANGER’S CHILD.

The mail carrier recently delivered an Amazon package containing a gift from jamessal, a copy of The Stranger’s Child, by Alan Hollinghurst. I set it aside till I had finished my latest copyediting slog, as a reward, and now that I’ve started it I can immediately see why Jim called Hollinghurst “a world-class writer” and wanted to share the book with me. Here are the first two paragraphs (virtually the entire first page):

She’d been lying in the hammock reading poetry for over an hour. It wasn’t easy: she was thinking all the while about George coming back with Cecil, and she kept sliding down, in small half-willing surrenders, till she was in a heap, with the book held tiringly above her face. Now the light was going, and the words began to hide among themselves on the page. She wanted to get a look at Cecil, to drink him in for a minute before he saw her, and was introduced, and asked her what she was reading. But he must have missed his train, or at least his connection: she saw him pacing the long platform at Harrow and Wealdstone, and rather regretting he’d come. Five minutes later, as the sunset sky turned pink above the rockery, it began to seem possible that something worse had happened. With sudden grave excitement she pictured the arrival of a telegram, and the news being passed round; imagined weeping pretty wildly; then saw herself describing the occasion to someone, many years later, though still without quite deciding what the news had been.
In the sitting-room the lamps were being lit, and through the open window she could hear her mother talking to Mrs. Kalbeck, who had come to tea, and who tended to stay, having no one to get back for. The glow across the path made the garden suddenly lonelier. Daphne slipped out of the hammock, put on her shoes, and forgot about her books. She started towards the house, but something in the time of day held her, with its hint of a mystery she had so far overlooked: it drew her down the lawn, past the rockery, where the pond that reflected the trees in silhouette had grown as deep as the white sky. It was the long still moment when the hedges and borders turned dusky and vague, but anything she looked at closely, a rose, a begonia, a glossy laurel leaf, seemed to give itself back to the day with a secret throb of colour.


The first thing that struck me was that with a few name changes, this could be straight out of To the Lighthouse; the ambience, the focus on psychology, and the gorgeous writing are very Woolfian. To take those items in order: the poetry in the hammock and the apparently substantial estate, with a rockery (UK for “rock garden”), pond, hedges, and borders, indicate a well-off family like the Ramsays, and the telegram and the lamps being lit suggest the pre-WWII world; the train of thought (poetry… where’s George? what’s Cecil like? … they’re very late… maybe they had an accident! he’s dead! … news, weeping, telling someone about it years from now…) is acutely observed and brilliantly presented, and the sentences! the language! “Now the light was going, and the words began to hide among themselves on the page”: utterly simple and yet I’ve never seen that particular event (so familiar to those of us who read in all surroundings and conditions of light) described so well and memorably; “weeping pretty wildly” is pitch-perfect; and the whole last passage starting “She started towards the house…” is downright Nabokovian in its deployment of sounds and syntax to render a complex perception, with the final “seemed to give itself back to the day with a secret throb of colour” somehow calling back to “the words began to hide among themselves on the page.” I’m hooked, and very much looking forward to the rest of the book. (Those of you who have read it should feel free to make general remarks, but please, no spoilers—I haven’t the faintest idea what the book is about yet!)
Addendum. Having posted this, I have to add this sentence from the Hollinghurst novel: “He spoke German nicely, keeping an amused pedantic eye on the slowly approaching end of his sentences.”

Comments

  1. Perfect purple prose for those who take their sighing seriously. Set it to music and you’ve got a soap opera voice-over. I keep hearing “Oh Daphne!” echoing from the page.

  2. No, it’s not Hemingway, if that’s the kind of thing you prefer.

  3. The interesting question, Hozo, is what sets this ambitious, highly descriptive, musical prose apart from purple imitations. Hat took a pretty damn good shot at answering it, too, by pointing out that it’s not just ornate; it’s immensely informative (i’ve been reading pages twice, three times over with pleasure), and the insights are delivered in prose as plain as you could desire (“the words began to hide among themselves on the the page”). You won’t find a single serious critic who doesn’t admire Hollinghurst’s prose — or, at least, I doubt you will (I’ve read a lot of reviews) — which is strange, considering how easy it is to crack wise.

  4. It’s overwrought, tedious, affected and full of cliched atmospherics. If you like your literature dreary and dripping with point of consciousness slurry then you’ve come to the right place. Puts me in mind of the British pre-Raphaelite painters. For all the stimuli their literal realism thought to provoke, what lingers is a maudlin sentimentality uniquely bereft of pictorial heft.

  5. uniquely bereft of pictorial heft.
    Lovely internal rhyming there, mate. Just the sort of thing I look for in good prose. Puts me in mind of the Hallmark school of art.

  6. uniquely bereft of pictorial heft.
    Lovely internal rhyming there, mate. Just the sort of thing I look for in good prose. Puts me in mind of the Hallmark school of art.

  7. Boy, I hope for your sake you have a better awareness of literature than you do of art.

  8. Caroline says:

    Hat, you are going to love the “adverb game” they play– very briefly!– in the second section. A rather minor character uses a word that mystifies all the other players and hints delightfully at a vast interiority otherwise basically unplumbed by Hollinghurst.

  9. AJP-
    I’d recommend Pink Chablis for the above passage. I prefer sake for more invigorating stimulation.
    ps apropos the nonsensical “aware”!
    http://www.youtube.com/watch?v=Vi5JyCYKZws

  10. Aware isn’t nonsense; it means watchful, vigilant, cautious, on one’s guard, informed, cognizant, conscious, sensible, generally concerned & well informed, esp. with respect to a particular issue or field. In your case, you aren’t well informed about art – either that or your loathing (hatred, detestation, abhorrence, abomination, execration, antipathy, dislike, hostility, animosity, ill feeling, bad feeling, malice, animus, enmity, aversion, repugnance) is distorting (twisting, warping, contorting, buckling, deforming, disfiguring) your judgement.
    I’ll give you one more reply, but I have to watch television with my dog now.

  11. Oh dear.
    I feel so stupid.
    My train of thought can be summed up as “yes, that sounds rather nice, I guess”.

  12. marie-lucie says:

    Since we are looking at (quasi-)synonyms, I think that (except for the rhyme) devoid would be better than bereft here. Bereft, originally a form of the verb bereave, implies that something which used to exist has been lost. I don’t think that is the meaning intended here.
    As for the pictorial heft of the passage, I think there is plenty of that. I can very clearly see the scene in my mind.

  13. An informed judgement has no need to tether loathsomeness to its guidon. In all things moderation, even those which moderate towards the extreme.

  14. jamessal says:

    Are you talking to yourself?

  15. Hozo is simply upholding the “Your favorite band sucks!” ethos of the lower strata of the internet.

  16. How unlike you to play the high Hat sir. Les bas fonds have been known to produce mellifluous strains. Leaving aside the pricklish ivories to the towered sycophants and their “interiority” complex.

  17. Ages ago I read Hollinghurst’s first novel, The Swimming Pool Library. It is a gay hothouse perverted egocentric intellectual ’80s kind of book. I simply adored it, but found his next novel, The Folding Star, unreadable.
    I have finally gotten all my books back from storage into my apartment, and am now sorting them into the shelves. I’ll look into Library again when I run across it. The ’80s are over now, so I suspect I won’t read far.

  18. dearieme says:

    I likee: the only false note (is it?) is “reading poetry”. What, any old poetry? Or perhaps that’s the point: Daphne perhaps thinks of poetry as a rather undifferentiated class.

  19. Bathrobe says:

    宝三 (is that how you would write your name?), given the growing hoity-toityness of your posts, what kind of prose literature would you suggest that the discriminating reader should read?

  20. the only false note (is it?) is “reading poetry”. What, any old poetry?
    No, that’s a setup. The question of exactly what poetry she reads is raised shortly thereafter.

  21. Confidential to Hozo: To quote Professor Calvin Leslie Lewis, A Handbook of American Speech (Scott, Foresman and Company, 1916), p. 29: “If one does not wish to be thought a fault-finder, a scold, a dolt, or a pig he should not talk like one.”

  22. Now, now; come, come, people. Next we will be hearing that this is the sort of thing that those (or that only those, depending) who like this sort of thing will like. (As indeed it is.) The above quotation is not going to be to everyone’s taste, and defenses of taste (to misquote Frye) are usually intelligible only to those well within the defenses. (Add remark about defensiveness here.)
    But there’s no need to go as far as Hemingway (“Sentence fragments. Good device. Especially if not overused.”) to find contrasts with this. I can scarcely imagine Hat himself, for example, writing anything like this. (That’s not a criticism.)
    In short, let’s not snipe at one another. There are many, many targets which deserve it far more.

  23. No, Hozo definitely deserves it; it’s entirely possible (indeed, almost everyone else seems to manage it without much effort) to disagree about style, or anything else, without coming off as a supercilious jerk, and Hozo has been doing the latter with dispiriting consistency.

  24. jamessal says:

    Hemingway (“Sentence fragments. Good device. Especially if not overused.”)
    As I’m sure John C. knows, Hemingway also wrote a lot of paratactic (no subordinators), polysyndetic (“and… and… but… and… and… but… but…”) page-long sentences, too, a few of them expressing perfectly mawkish machismo. Not that he wasn’t often wonderful, too, of course.
    In short, let’s not snipe at one another.
    John, much as I admire your wish to keep the peace, we’re simply not sniping at one another; we’re telling Hozo to behave or go away. It’s not like he said, “This passage isn’t for me, and these are the reasons…”; no, he took a big florid dump on it right after Hat had expressed his enthusiasm, and that’s simply assholic.

  25. jamessal says:

    Apparently, if I’d waited another minute, I could have just said ditto.

  26. Enough of HoJo, it’s boring. I’d be interested to know if Jamessal has read either The Swimming-Pool Library or The Line Of Beauty. I haven’t, although I saw an adaptation of The Line Of Beauty on television last year, and it was good. I find this sort of thing interesting partly because I grew up in England and am probably about Hollinghurst’s age – I lived fairly near Wealdstone until I was seven, in those days I thought it was called Wheelstone – so I’m slightly surprised the subject matter (and these books seem to be very much about their subject matter, pre-aids gay England) would interest Language & Jamessal.

  27. jamessal says:

    No, this my first Hollinghurst, though I’ve also downloaded The Line of Beauty. As for the subject matter, that’s about the last of my concerns when considering which books to read, the first being the writing.

  28. jamessal says:

    For example, supercilious pederasts, dreaming seven-year-olds, and men who suck on stones in a very precise fashion don’t all draw me to the edge of my seat, but… I think you can tell where I’m going with this.

  29. J.W. Brewer says:

    Are we going to get one of those semi-incomprehensible British newspaper headlines out of this book? (E.g., “Wealdstone Supercilious Pederast Shock Horror” – or is that “paederast” even in headlines?)

  30. jamessal says:

    That was funny, but for the sake of clarity I was referring to Lolita, Alice in Wonderland, and Molloy.

  31. jamessal says:

    Speaking of pederasty, I had known etymologically that it referred to a love only of young boys, but had also thought that the meaning had been extended to all forms of underage molestation decades ago. Am I just wrong the second part? The dictionaries I’m checking seem to be sticking with the boys.

  32. J.W. Brewer says:

    Not having read Hollinghurst, I had supposed his plots might feature all three of those character types . . . Google has “about 22″ results for “humbert pedophile” (in that order, with no intervening words) but zero for “humbert pederast,” which is consistent with my own sense that pederast remains boy-specific for most AmEng speakers whereas pedophile covers those attracted to underage persons of either sex (w/o getting into the pedophile/ephebophile distinction mooted in some quarters of the media a few years back).

  33. Bathrobe says:

    I was using hoity toity in the archaic sense of “given to frivolity, silliness or riotousness”.

  34. my own sense that pederast remains boy-specific for most AmEng speakers whereas pedophile covers those attracted to underage persons of either sex
    That’s certainly my usage.

  35. marie-lucie says:

    In French, pédéraste means ‘homosexual male’, without reference to the age of the man’s preferred choices. This word is very derogatory, especially its abbreviation pédé, and perhaps for this reason the English word gay seems to be preferred nowadays.
    Bathrobe: I did not know this meaning of hoity toity. Was the current meaning influenced or even caused by the similarity of the syllable hoi with the word high ?

  36. Bathrobe says:

    m-l, I cannot tell a lie, I got it from here. And yes, it appears that the change in meaning came about as you said.
    I was groping for a word to describe Hozo’s descent from scorn to silliness and this one somehow came up, although I didn’t fully realise the big historical change in its meaning.

  37. Bathrobe says:

    m-l, I cannot tell a lie, I got it from here. And yes, it appears that the change in meaning came about as you said.
    I was groping for a word to describe Hozo’s descent from scorn to silliness and this one somehow came up, although I didn’t fully realise the big historical change in its meaning.

  38. jamessal says:

    That’s certainly my usage.
    And mine now, as well — thanks Hat and J.W.! Good money says, however, that the meaning will be extended even in carefully edited prose within a century or so. I’m reading The English Language: A Historical Introduction by Charles Barber; it gives the best quick explanation I’ve read of natural linguistic evolution — from assimilation (“the changing of a sound under the influence of a neighbouring one”) to analogy (“the process of inventing a new element in conformity with some part of the language system that you already know”) — and, as Barber says, “The rarer a word is, the more likely it is to be affected by analogy.”

  39. J.W. Brewer says:

    Could be, although it might also be the case that pederast will be completely crowded out by pedophile, which has the advantage (in some circles) of implying that adult-youth sexual interaction is worthy of strong condemnation without implying any condemnation whatsoever of male-male homosexuality more broadly. “Pederast” in fact perhaps already has a semi-archaic feel (although not so much as e.g. “sodomite”), whereas “pedophile” sounds all legal/medical and thus up-to-date.

  40. jamessal says:

    “Pederast” in fact perhaps already has a semi-archaic feel
    Huh, I wouldn’t have thought it sounded archaic. To me it just sounds harsher than pedophile (even though, as you say, there’s a different kind of clinical harshness to that word). The last vowel is an ash, so you end up snarling, and it has the word ass built into it, so you have that image, before you finish not on a smooth l but a plosive. Just a harsher sounding word.

  41. Jim: analogy (“the process of inventing a new element in conformity with some part of the language system that you already know”) — and, as Barber says, “The rarer a word is, the more likely it is to be affected by analogy.”
    What does “affected” by analogy mean ? Let’s take a word X that already exists and is rare. How does analogy now come into the picture, and what aspect of X does it “affect”, and why ? Who would be trying to do what with X by employing analogy ? Where is the “invention” ?
    I’m wondering whether what Barber means to say would be more clearly expressed by: “The rarer a word is, the more likely it is to have arisen by analogy”.
    What does Barber have to say about clarity in English ?

  42. it has the word ass built into it, so you have that image, before you finish not on a smooth / but a plosive.
    “Assplosive” would be a useful word, if only I could think of a use for it. Perhaps in connection with lurid butt bombs.

  43. What does “affected” by analogy mean ?
    “Analogy” here is something of a specialized term, and since apparently I didn’t provide enough context before, I’m pasting below Barber’s whole section on linguistic evolution (I inserted all the paragraphs, and bolded the most important terms, but it would have taken me all night to italicize everything that’s italicized in the book):

    Mechanisms of linguistic change
    All living languages undergo changes analogous to those we have just seen exemplified in English. What causes such changes? There is no single answer to this question: changes in a language are of various kinds, and there seem to be various reasons for them.
    The changes that have caused the most disagreement are those in pronunciation. We have various sources of evidence for the pronunciations of earlier times, such as the spellings, the treatment of words borrowed from other languages or borrowed by them, the descriptions of contemporary grammarians and spelling-reformers, and the modern pronunciations in all the languages and dialects concerned. From the middle of the sixteenth century, there are in England writers who attempt to describe the position of the speech-organs for the production of English phonemes, and who invent what are in effect systems of phonetic symbols. These various kinds of evidence, combined with a knowledge of the mechanisms of speech-production, can often give us a very good idea of the pronunciation of an earlier age, though absolute certainty is never possible.
    When we study the pronunciation of a language over any period of a few generations or more, we find there are always large-scale regularities in the changes: for example, over a certain period of time, just about all the long [a:] vowels in a language may change into long [e:] vowels, or all the [b] consonants in a certain position (for example at the end of a word) may change into [p] consonants. Such regular changes are often called sound laws. There are no universal sound laws (even though sound laws often reflect universal tendencies), but simply particular sound laws for one given language (or dialect) at one given period. We must not think of a sound law, however, as a sudden change which immediately affects all the words concerned. If [b] changes to [p] in a given language, the change may first appear in words which are frequently used, and gradually spread through the rest of the vocabulary. Indeed, the sound law may cease to operate before all the relevant words have been affected, so that a few are left with the earlier pronunciation.
    One cause which has been suggested for changes in pronunciation is geographic and climatic, for example that people living in mountain country are subject to certain changes in pronunciation compared to plainsmen, but the evidence for this is unconvincing. Other people have suggested biological and racial factors: it has been said, for example, that races with thick lips have difficulty in producing certain speech-sounds. Once again, no really convincing evidence has been produced: and if a child of any racial origins is brought up from birth in a normal English-speaking family, it will grow up speaking English just like a native. Moreover, the theory would obviously be most useful for explaining changes in a language when it is adopted by one people from another. But in these circumstances the theory is unnecessary: the influence of one language on another is quite enough to explain such changes, without racial characteristics being invoked.
    During childhood, we learn our mother tongue very thoroughly, and acquire a whole set of speech habits which become second nature to us. If later we learn a foreign language, we inevitably carry over some of these speech habits into it, and so do not speak it exactly like a native. For example, we have seen that in most phonetic contexts the English /p/ phoneme is pronounced with a following aspiration, producing a kind of [ph] sound, and the same is in fact true of the English /t/ and /k/ phonemes. But it is not true of the similar phonemes in French or Italian, where the voiceless plosives are pronounced without any following aspiration. Many English speakers of French and Italian, even competent ones, carry over their aspirated voiceless plosives into those languages, and this is one of many features that make them sound foreign to native speakers. In bilingual situations, therefore, the second language tends to be modified. Such modifications may not persist: an isolated Polish or Pakistani immigrant to Britain will usually have grandchildren who speak English like natives, because the influence of the general speech environment (peer-group, school, work) is stronger than that of the home. But if a large and closely-knit group of people adopt a new language, then the modifications that they make in it may persist among their descendants, even if the latter no longer speak the original language that caused the changes. This can be seen in Wales, where the influence of Welsh has affected the pronunciation of English, and the very characteristic intonation-patterns of Welsh English have been carried over from Welsh, even among those who no longer speak it. Many historical changes may have been due to a linguistic substratum of this kind: a conquering minority that imposed its language on a conquered population must often have had its language modified by its victims.
    It is also possible that fashion plays a part in the process of change. It certainly plays a part in the spread of change: one person imitates another, and people with the most prestige are most likely to be imitated, so that a change that takes place in one social group may be imitated (more or less accurately) by speakers in another group. When a social group goes up or down in the world, its pronunciation may gain or lose prestige. It is said that, after the Russian Revolution of 1917, the upper-class pronunciation of Russian, which had formerly been considered desirable, became on the contrary an undesirable kind of accent to have, so that people tried to disguise it. Some of the changes in accepted English pronunciation in the seventeenth and eighteenth centuries have been shown to consist in the replacement of one style of pronunciation by another style already existing, and it is likely that such substitutions were a result of the great social changes of the period: the increased power and wealth of the middle classes, and their steady infiltration upwards into the ranks of the landed gentry, probably carried elements of middle-class pronunciation into upper-class speech.
    Besides spreading changes that have already taken place, fashion may actually cause changes in pronunciation. The important thing about a fashion is that it’s exclusive: as soon as the fashion has penetrated to a less prestigious social group, it’s time to move on. This can be seen in clothes: fashionable people may find it flattering to be imitated, but as soon as the new fashion has really caught on, they need to change to something else, to mark themselves off as different. It may be the same with language, for social groups use characteristic styles of language to mark themselves off from other groups. A group with high prestige may find that its style of speech is being imitated by other groups, and then its members may (perhaps unconsciously) begin to change it, perhaps by exaggerating its distinguishing characteristics.
    Another suggested cause for changes in pronunciation is the fact that children grow. The vocal organs of children, it is argued, are a different size from those of adults: they learn to mimic the noises that their parents make, but on what is in effect a different instrument; as they grow up, they go on moving their vocal organs in the same way, but the sounds that they produce are now different, because the organs are changed. There are two serious objections to this theory, however. The first is that, if this were indeed a major cause of phonological change, we should then expect all changes of pronunciation to be in the same direction, irrespective of language or period, and this is certainly not the case. The second objection is that humans appear to have the capacity, probably innate, to allow for different sizes of vocal tract when they interpret speech: the voices of a young child, of a woman, and of a deep-voiced man all have different pitch-levels, but this does not cause problems of understanding. When the man utters the vowel [ɪ] he may in fact produce the same acoustic signal as the woman when she says [e], or the child when it says [æ], but we have an inbuilt mechanism that allows us to interpret their vowels correctly, provided we hear them as part of a longer utterance. This fact makes it unlikely that growth of the vocal tract leads to changes in pronunciation.
    A less specific variant of the argument is that the imitation of children is imperfect: they copy their parents’ speech, but never reproduce it exactly. This is true, but it is also true that such deviations from adult speech are usually corrected in later childhood. Perhaps it is more significant that even adults show a certain amount of random variation in their pronunciation of a given phoneme, even if the phonetic context is kept unchanged. This, however, cannot explain changes in pronunciation unless it can be shown that there is some systematic trend in the failures of imitation: if they are merely random deviations they will cancel one another out and there will be no net change in the language. For some of these random variations to be selected at the expense of others, there must be further forces at work.
    One such force which is often invoked is the principle of ease, or minimization of effort. We all try to economize energy in our actions, it is argued, so we tend to take short cuts in the movements of our speech-organs, to replace movements calling for great accuracy or energy by less demanding ones, to omit sounds if they are not essential for understanding, and so on. Such changes increase the efficiency of the language as a communication-system, and are undoubtedly a factor in linguistic change, though we have to add that what seems easy or difficult to a speaker will depend on the particular language that has been learnt. Suppose we have a sequence of three sounds in which the first and the third are voiced, while the middle one is voiceless: the speaker has to carry out the operation of switching off voice before the second sound and then switching it on again before the third. An economy of effort could be obtained by omitting these two operations and allowing the voice to continue through all three sounds. Such a change would be seen if the pronunciation of fussy were changed to fuzzy, the voiceless /s/ being replaced by the voiced /z/ between the two vowels. Changes of this kind are common in the history of language, but nevertheless we cannot lay it down as a universal rule that fuzzy is easier to pronounce than fussy. In Swedish, for example, there is no /z/ phoneme, and Swedes who learn English find it difficult to say fuzzy, which they often mispronounce as fussy. For them, plainly, fussy is the easier of the two pronunciations, because it accords better with the sound-system of their own language.
    The change from fussy to fuzzy would be an example of assimilation, which is a very common kind of change. Assimilation is the changing of a sound under the influence of a neighbouring one. For example, the word scant was once skamt, but the /m/ has been changed to /n/ under the influence of the following /t/. Greater efficiency has hereby been achieved, because /n/ and /t/ are articulated in the same place (with the tip of the tongue against the teeth-ridge), whereas /m/ is articulated elsewhere (with the two lips). So the place of articulation of the nasal consonant has been changed to conform with that of the following plosive. A more recent example of the same kind of thing is the common pronunciation of football as foopball. Sometimes it is the second of the two sounds that is changed by the assimilation. This can be seen in some changes that have taken place in English under the influence of /w/: until about 1700, words like swan and wash rhymed with words like man and rash; the change in the vowel of swan and wash has given it the lip-rounding and the retracted tongue-position of the /w/, and so economized in effort.
    Assimilation is not the only way in which we change our pronunciation in order to increase efficiency. It is very common for consonants to be lost at the end of a word: in Middle English, word-final /-n/ was often lost in unstressed syllables, so that baken ‘to bake’ changed from /’ba:kən/ to /’ba:kə/, and later to /ba:k/. Consonant-clusters are often simplified. At one time there was a /t/ in words like castle and Christmas, and an initial /k/ in words like knight and know. Sometimes a whole syllable is dropped out when two successive syllables begin with the same consonant (haplology): a recent example is temporary, which in Britain is often pronounced as if it were tempory.
    On the other hand, ease of pronunciation can lead to an extra phoneme being inserted in a word: in Old English, our word thunder was þunor, with no d. By normal development, þunor would have become *thunner, not thunder, but at some stage a /d/ has been inserted in the pronunciation. Spellings with d are first found in the thirteenth century, and are completely normal by the sixteenth. Why was a /d/ inserted in the word? Probably because the pronunciation thunder actually calls for less precise movements of the speech-organs. The /d/ arose from a slight mistiming in the transition from the nasal /n/ to the following phoneme (which was probably a syllabic /r/ rather than a vowel). This transition calls for two simultaneous movements of the speech-organs: (1) the nasal passages are closed by the raising of the soft palate, and (2) the tongue is moved away from the teeth to unblock the mouth-passage. If the two movements are not carried out simultaneously, but the nasal passages are closed before the tongue moves, a /d/ will be heard between the /n/ and the following phoneme, as the stop is released. Similar mistimings produced the /b/ in the middle of the words thimble and bramble (Old English þymel, brēmel). Sometimes, too, ease of pronunciation apparently leads us to reverse the order of two phonemes in a word (metathesis): this has happened in the words wasp and burn, which by regular development would have been waps and brin or bren.
    The changes produced in pursuit of efficiency can often be tolerated, because a language always provides more signals than the absolute minimum necessary for the transmission of the message, to give a margin of safety: like all good communication-systems, human language has built in to it a considerable amount of redundancy. But there is a limit to this toleration: the necessities of communication, the urgent needs of humans as users of language, provide a counterforce to the principle of minimum effort. If, through excessive economy of effort, an utterance is not understood, or is misunderstood, the speaker is obliged to repeat it or recast it, making more effort. The necessities of communication, moreover, may be responsible for the selection of some of the random variations of a phoneme rather than others, so that a change in pronunciation occurs in a certain direction. This direction may be chosen because it makes the sound inherently more audible: for example, open nasal vowels seem to be more distinctive in quality than close ones, and in languages which have such vowels it is not uncommon for a nasal [e] to develop into a nasal [a].
    In considering such changes, however, we cannot look at the isolated phoneme: we have to consider the sound-system of the language as a whole. The ‘safeness’ or otherwise of a phoneme for communicative purposes does not depend solely on its own inherent distinctiveness: it depends also on the other phonemes in the language with which it can be contrasted, and the likelihood that it may be confused with them. Let us imagine that in the vowel-system of a language there is a short [e], as in bet (see for example the vowel-diagram in Figure 4, p. 14 above); in one direction from it there is a short [æ] (as in bat), and in another direction a short [a] (as in the first syllable of about); but in the upward (closer) direction there is no short vowel, no kind of short [ɪ] for example. Suppose now that random variations occur in speakers’ pronunciations of these three vowels. When the variations of [e] go too far in the direction of [æ] or [a], the speaker will be forced to correct them, to avoid misunderstanding. But when the variations are in the direction of [ɪ], there is no such necessity for checking or correction. The result will be a shift in the centre of gravity of the [e], which will drift up towards [ɪ]. Moreover, the movement of [e] towards [ɪ] will leave more scope for variations in [æ], which may tend to drift up towards [e]. In this way, a whole chain of vowel-changes may take place.
    In this example I have assumed that the contrast between the three vowels is important enough in the functioning of the language for speakers to resist any changes which threaten this contrast. This will be the case if large numbers of words are distinguished from one another by these vowels, in other words if the contrast between them does a lot of work in the language. The functional load carried by a contrast is a major factor when speakers decide (unconsciously) whether to let a change take place or not. There may be forces in the system making for the amalgamation of two phonemes, and if there are very few words in the language which will be confused with one another as a result then there will not be much resistance to the change; but if serious confusion will be caused by the amalgamation it will be resisted more strongly, and perhaps be prevented.
    This does not mean, on the other hand, that a phoneme with a small functional load will necessarily be thrown out of the system, either by being lost or by being amalgamated with another phoneme. It also depends on the degree of effort required to retain the phoneme, which may be quite small. For example, the contrast in English between the voiced /ð/ and the voiceless /þ/ phonemes carries a very small load; there are a few pairs of words that are distinguished from one another solely by this difference, like wreathe and wreath, and mouth (verb) and mouth (noun); but in practice the distinction between the two phonemes is of very small importance, and it would cause no great inconvenience if they were amalgamated, for example by both evolving into some third, different, phoneme. On the other hand, it takes very little effort to retain the distinction between them. They belong to a whole series of voiced and voiceless fricatives (/v/ and /f/, /z/ and /s/, /ʒ/ and /ʃ/), and so fall into a familiar pattern; and if we abolished the distinction between them we should not economize in the number of types of contrast that we made; we should still have to distinguish fricatives from other types of consonant, and between voiced and voiceless fricatives.
    The stability of /ð/ and /þ/ thus results from the fact that they are, in Andre Martinet’s terminology, ‘well integrated’ in the consonant system of English. An even better integrated group of consonants in present-day English is the following:
    Voiceless plosives /p/ /t/ /k/
    Voiced plosives /b/ /d/ /g/
    Nasals /m/ /n/ /ŋ/
    Each of these three series uses the same places of articulation: the two lips pressed together for /p/, /b/, /m/; the tip of the tongue pressed against the teeth-ridge for /t/, /d/, /n/; the back of the tongue pressed up against the soft palate for /k/, /g/, /ŋ/. So, using only three articulatory positions, and three distinctive articulatory features (plosiveness, nasality, voice), we get no fewer than nine distinct phonemes. This group is very stable, because the loss of any one of the nine would produce negligible economy in the system: if, say, /ŋ/ were to disappear, we should still have to be able to produce nasality for /m/ and /n/, and we should still have to be able to articulate with the back of the tongue against the soft palate for /g/ and /k/. So even if /ŋ/ carried a very small load in the language we should still be unlikely to get rid of it. For the same reason, if there were a hole in the pattern, it would stand a good chance in time of getting filled. If there were no /ŋ/ in present-day English, but there was some other consonant which was not very well integrated in any sub-system, then any variations in this consonant that moved it in the direction of [ŋ] would tend to be accepted, because they would represent an ‘easier’ pronunciation — easier, that is, in terms of the economy (and therefore efficiency) of the system as a whole.
    Changes in morphology, syntax, vocabulary, and word-meaning, while they can be complicated enough, are less puzzling than changes in pronunciation. Many of the same causes can be seen at work. The influence of other languages, for example, is very obvious: nations with high commercial, political, and cultural prestige tend to influence their neighbours: for centuries, French influenced all the languages of Europe, while today the influence of the English language is penetrating all over the world, largely because of the power and prestige of the United States. This influence is strongest in the field of vocabulary, but one language can also influence the morphology and syntax of another. Such influence may occur if languages in a given area are in intimate contact over an extended period, and also when a religion spreads and its sacred books are translated: in the Old English period there were many translations from the Latin, and there is some evidence that Latin syntax influenced the structure of Old English.
    In the realm of vocabulary and meaning, the influence of general social and cultural change is obvious. As society changes, there are new things that need new names: physical objects, institutions, sets of attitudes, values, concepts; and new words are produced to handle them (or existing words are given new meanings). Sentimentality, classicism, wave mechanics, parliaments, post-Impressionism, privatization — these are human inventions just as much as steam engines or aircraft or nylon: and people inevitably invented names for them. Moreover, because the world is constantly changing, many words insensibly change their meanings. It is particularly easy to overlook shifts of meaning in words that refer to values or to complexes of attitudes: for example, in Shakespeare’s day the adjective gentle meant a good deal more than ‘kind, sweet-natured, mild, not violent’, for it referred to high birth as well as to moral qualities, and had a whole social theory behind it.
    As in pronunciation, so at the other levels of language, we see the constant conflict between the principle of minimum effort and the demands of communication. Minimization of effort is seen in the way words are often shortened, as when public house becomes pub, or television becomes telly, and also in the laconic and elliptical expressions that we often use in colloquial and intimate discourse. But if economy of this kind goes too far, some kind of compensating action may be taken, as when in Early Middle English the word ea was replaced by the French loan-word river, and in the seventeenth century the bird called the pie was expanded to the magpie. In such ways, the redundancy which has been removed from the language by shortenings may be reinserted by lengthenings.
    There is also interplay between the needs of the users and the inherent tendencies of the language-system itself. One way in which the language-system promotes change, especially in grammar, is through the operation of analogy, which also tends to produce economy. Analogy is seen at work when children are learning their language. A child learns pairs like dog/dogs, bed/beds, bag/bugs, and so on. Then it learns a new word, say plug, and quite correctly forms the plural plugs from it, on analogy with these other pairs. Analogy, then, is the process of inventing a new element in conformity with some part of the language system that you already know. The way in which analogy can lead to change is seen when the child learns words like man and mouse, and forms the analogical plurals mans and mouses. Ultimately such childish errors are usually corrected, but analogical formations also take place in adult speech, and quite often persist and become accepted. In Old English there were many different ways of putting a noun into the plural: for example, stān ‘stone’, stānas ‘stones’; word ‘word’, word ‘words’; scip ‘ship’, scipu ‘ships’; synn ‘sin’, synna ‘sins’; tunge ‘tongue’, tungan ‘tongues’; bēo ‘bee’, bēon ‘bees’; boc ‘book’, bēc ‘books’; lamb ‘lamb’, lambru ‘lambs’. The form stānas has developed quite regularly into our plural stones, but, sometime during the past thousand years, all the others have changed their plural ending to the -(e)s type, by analogy with the many nouns like stone. The rarer a word is, the more likely it is to be affected by analogy. The unusual noun-plural forms in present-day English, which are the ones that have managed to resist the analogy of the plural in -(e)s, are mostly very common words, like men, feet, and children, or at any rate are words which were very common a few centuries ago, like geese and oxen.
    Charles Barber, The English Language: A Historical Introduction (Canto) (Kindle Locations 504-647). Kindle Edition.

    I only elided one example of old English — sin — because apparently the way ’twas spelt in the olden days is considered “questionable content.” [Fixed—LH]

  44. it would have taken me all night to italicize everything that’s italicized in the book
    This is a clear case of TMI (Too Much Italicization).

  45. If you take this opportunity to make an admittedly funny pun, Stu, instead of reading the material that will answer your above questions — and was a huge pain in the ass to paste (Kindle doesn’t make it easy) — I will never forgive you.

  46. It still doesn’t make sense. Even in the context you quote, the sentence “the rarer a word is, the more likely it is to be affected by analogy” brings my thought processes to a screeching halt.
    All of the words given as examples are common words:

    In Old English there were many different ways of putting a noun into the plural: for example, stain ‘stone’, stanas ‘stones’; word ‘word’, word ‘words’; scip ‘ship’, scipu ‘ships’; … ‘; tunge ‘tongue’, tungan ‘tongues’; beo ‘bee’, Non ‘bees’; boc ‘book’, bec ‘books’; lamb ‘lamb’, lambru ‘lambs’. The form stanas has developed quite regularly into our plural stones, but, sometime during the past thousand years, all the others have changed their plural ending to the -(e)s type, by analogy with the many nouns like stone. The rarer a word is, the more likely it is to be affected by analogy. The unusual noun-plural forms in present-day English, which are the ones that have managed to resist the analogy of the plural in -(e)s, are mostly very common words, like men, feet, and children, or at any rate are words which were very common a few centuries ago, like geese and oxen.

    No example is given of a rare word, so no example is given of a rare word “affected by analogy”. I am now more certain than before that Barber has not said what he has to say in such a way that it can be understood – or else there is a misprint. If anything, his examples lead me to think: “The rarer a word is, the more unlikely it is to be affected by analogy”.

  47. Calm down ! If I had read the entire article before making that silly pun, I would have forgotten the pun. When I scratch my nose before replying to your long comment, that is not the same as thumbing my nose at it.

  48. The old plural of “beo” meaning “bee” was “Non”?

  49. First, I’m perfectly calm — I was making a joke. Second, it makes perfect sense. Common words are used so commonly that if someone analogizes one incorrectly, it isn’t likely to make much dent in the language as a whole. If a word is rarely used, on the other hand, and someone thinks they can induce its function or its meaning or whatever from the rest of their knowledge of the language, someone else might pick up on that mistake, and, rare as the word is, you’ll soon have a competing variant. That’s what he’s saying.

  50. The old plural of “beo” meaning “bee” was “Non”?
    Good catch. No, it was beon with a macron over the e. The be, with its long e, must have gotten stuck in the series of tubes known as the internet — someone get the snake!

  51. I really hope there aren’t too many of those.

  52. I am now more certain than before that Barber has not said what he has to say in such a way that it can be understood
    Setting aside the fact that I understood it, you have to understand that this is not a textbook about linguistic change but a history of English with a crash course in linguistic change at the beginning. I happen to think the crash course kicks ass. He can’t give examples of everything, but he can give you enough sense of how language works to work an idea out yourself when he needs to be more terse.

  53. marie-lucie says:

    That’s quite a job, jamessal! There must have been several phonetic symbols that did not go through intact, like the one for the nasal consonant usually represented as ng, which shows up here as q, and the ones for the two types of th (voiced in “with” and voiceless in “teeth”). But these are small details which do not really detract from the information in the whole chapter.
    Rare words and analogy: the idea is that if a word with an unusual secondary form (of plural for English nouns, past tense for English verbs) is rare, people will forget the unusual, irregular form and replace it with one in conformity (or analogy) with the most general type: thus for instance cows instead of the older plural kine. Obviously this is a tendency rather than a rule, as it depends on how common or rare the word is in a particular context or group of people: it could be that kine was replaced by cows but geese did not become gooses because only rural people lived with many cows, while geese were raised, like chickens, in small numbers even in urban areas. But it is hard to invoke rarity with replacements such as helped instead of older holp (there are many such examples with verbs which nowadays have regular past tenses). In the case of bees instead of beon, which would have ended up with a pronunciation been if it had not been replaced, the change may have been due to potential confusion with bean.

  54. jamessal says:

    So, M.L., re analogy, it’s strictly a question of form and not, as I had said, “function or meaning or whatever”?

  55. NOW I understand what is meant, thanks to marie-lucie ! “Unusual, irregular form” says what needs to be said here: irregularity within a word’s grammatical forms. “Rare word” says only: word occurring infrequently.
    When a special linguistic aspect of forms is being discussed, it is confusing that rare plural forms of a word are referred to as rare words, rather than as unusual, irregular plural forms. In the present context, “rare” is a poor substitute for “unusual and irregular”.
    I don’t especially care about Barber and his formulations. What interests me is this particular example of how easy or hard a reader can find it to understand a text, depending on exactly which words the author of the text chooses. Who was it who said something like: “good prose is not easier to write than good poetry” ?

  56. This is all very interesting. Thanks, Jim & m-l. It’s too bad we lost kine.
    the change may have been due to potential confusion with bean
    In Norwegian, “beans” and “farmers” – bønner and bønder – are pronounced almost exactly the same, but since everyone is aware of the similarity, there is never any confusion.
    Jimsal: As for the subject matter, that’s about the last of my concerns when considering which books to read, the first being the writing.
    Not that I have a completely tin ear or anything, but, fiction or non-, the subject matter is usually a fairly high consideration for me.

  57. In Norwegian, “beans” and “farmers” – bønner and bønder – are pronounced almost exactly the same, but since everyone is aware of the similarity, there is never any confusion.
    I suppose that many languages have homonyms. At one point when I was getting fed up with French homonyms, I thought over what the situation is in German and English. I was amazed to find that there are so many – and that they don’t usually cause any problem.
    When you look at it from a theoretical point of view, deploying notions like “homonym”, you might reach the conclusion that languages with homonyms must be a bitch to understand. But it seems that way only when you aren’t familiar with the words, as Crown says.
    This is further support for my claim that languages can be mastered only by forgetting what you learned about them.

  58. Grumbly, I think I have got this right. In modern German the pluralization of nouns follows many different patterns. Some nouns take -er with or without umlaut, others -e with or without unmlaut, others -en, ans so on. None of these would be called irregular. I gather that it was like that in Old English, but that for some reason one of the many patterns (namely the familiar one with -s) mostly won out over the others. There are a few nouns that never followed the crowd, and we sometimes call these vestiges of the older forms “irregular”.
    The point about “rare” is not that the irregular forms are rare, or anything like that. The point is that, on the whole, a noun that you used every day was more likely to resist the change than a noun that you didn’t use every day.
    Look again at what marie-lucie wrote:
    Rare words and analogy: the idea is that if a word with an unusual secondary form (of plural for English nouns, past tense for English verbs) is rare, people will forget the unusual, irregular form [...]

  59. The empty set is correct. Reread the last sentence in the Barber quote (which I have corrected by comparison with the source, adding italics and phonetic symbols; fortunately, I enjoy that sort of thing, which is why I am a good copyeditor): “The unusual noun-plural forms in present-day English, which are the ones that have managed to resist the analogy of the plural in -(e)s, are mostly very common words, like men, feet, and children…” The point is not that common words don’t get regularized (obviously, many do), but that it’s only common words that are likely to resist regularization.

  60. jamessal says:

    Thanks, Hat, for all the formatting, and M.L., for the more correct explanation — which, Grumbly, I’m sorry I couldn’t give you before. Apparently I didn’t understand it quite well enough. Good on you for prodding.
    Not that I have a completely tin ear or anything, but, fiction or non-, the subject matter is usually a fairly high consideration for me.
    If it’s non-fiction we’re talking, then I’m with you; it’s only fiction whose subject matter doesn’t concern me, and of course there’s nothing wrong with it concerning you more!

  61. No judgements are being made here. It just occurred to me that you can divide work up this way in many art forms (film, poetry, fine art). Even though it’s possible, it wouldn’t really work with, say, dance or architecture (assuming you consider architecture an art).

  62. marie-lucie says:

    LH: The point is not that common words don’t get regularized (obviously, many do), but that it’s only common words that are likely to resist regularization.
    Yes, this is the best way to put it. “Regularization” works by “analogical reformation” on the model of the most common (or “regular”) way of altering base words to add some grammatical meaning (eg forming plurals, tenses, etc). Small children do it at a certain stage in language acquisition, when they have mastered the regular pattern and apply it overall (eg saying hitted not hit) but before they figure out that those strange, counter-intuitive forms used by adults (eg hit instead of hitted, held instead of holded) do count as past forms, etc. For some words, even some adults can be confused about what the “proper” form should be, and use analogy to create new “irregular” forms on the model of existing patterns (eg brang or brung instead of brought, dove instead of dived, snuck instead of sneaked).
    The prime example of a very common word that has resisted regularization for centuries and even millennia in the Indo-European languages is the verb to be. In all these languages the verb corresponding to “to be” is unlike any other verb in being highly irregular, especially in the present tense. Not only is this verb extremely common, it does not have a definable, lexical meaning (“exist” is only rarely a possible replacement) but occurs mostly as another “function word”, and those are not often replaced by synonyms.

  63. jamessal says:

    assuming you consider architecture an art
    A safe assumption for all but an asshat!

  64. empty: In modern German the pluralization of nouns follows many different patterns. … None of these would be called irregular.
    That’s right. Some dictionaries provide a table of noun types (patterns) grouped according to how the nominative plural is formed. The types are numbered like T1, T2 etc. However, this is a layer of meta stuff that doesn’t help you much when learning, and must be forgotten later anyway.
    The patterns don’t help because you would need to memorize them and their numbers P1, P2 … and also to memorize which words belong to which patterns. You might as well just learn the words and their plurals, and dispense with the systematics. I have no clue how many types might be identified – around 10 ? Dunno and duccare.
    Perhaps somebody once imagined that such numbered noun-plural types might make dictionary entries slightly shorter, in addition to more scientific. The CD Duden doesn’t bother with this, but shows the genitive singular and nominative plural forms in an abbreviated fashion just following the lemma. Those are the forms that determine the whole declension (along with the nominative):

    Haus, das; -es, Häuser [mhd., ahd. hus, eigtl. = das Bedeckende, Umhüllende]:
    Matz, der; -es, -e u. Mätze [landsch. Kosef. des m. Vorn. Matthias] (fam. scherzh.): niedlicher kleiner Junge.
    Piep|matz, der (ugs.): [kleiner] Vogel:
    Spatz, der; -en, auch: -es, -en [mhd. spaz, spatze, Kosef. von mhd. spare, ahd. sparo, Sperling]:

    The Piep|matz lemma needs no details on the plural because the “|” tells you it’s a composite word, so for the pural you look up Matz. If there exist composite words with a plural form that is not the plural form of the last component (which I doubt), the plural details would be included with the lemmas for those composites.

  65. marie-lucie says:

    Irregular only means that there are few similar formations, and therefore no obvious rules for these formations. In English too there are several patterns for nouns and verbs, but in each category the list of irregulars is relatively short (especially for nouns), and for verbs, some formations are unique, often because other verbs which used to follow the same patterns have been regularized. In German, which has kept many of the old Germanic patterns, the list for each pattern is much longer, so a number of words can be classified under the same pattern. Of course the learner has to memorize the forms, but with separate lists for the various patterns it is possible, for instance, to find out and remember that a newly encountered word X (heard or read in just one of its forms) is listed in the same category and therefore follows the same pattern of plural or past formation as the already known words Y and Z, rather than some other potential pattern.

  66. Indeed, all of the German plural forms are justly described as irregular (not governed by rule) except -s, which is applied by rule when there is no usable analogy to a known plural. The regular form is not necessarily the most common form.
    In older Maori, most verbs ended in a consonant, and the regular passive ending was -ia. When final consonants were lost throughout the language, a formerly regular system broke down into a mass of irregularity: speakers simply had to remember whether the passive ending was -sia, -tia, -whia, and so on for about a dozen common endings and several more very rare ones. When Maori began to borrow verbs from English, it passivized them with -tia, for whatever reason, so -tia is now the regular ending (applied by rule when no other ending applies). But it is not and never has been the most common ending.
    In any case, the whole question of regularity conceals a deep mystery: why do languages bother with it? If people can learn 10,000 different nouns, they could learn 10,000 unique plurals for them — it’s not that much harder. But languages are not structured in that way, and nobody knows why.

  67. There is one hard and fast rule in German that I remember from school: “Dativ plural endet immer in -n“. It’s wonderful how it applies to nouns of all those different types, as well as any accompanying definite articles and (with or without article) adjectives.
    Of course, there turns out to be an exception: it doesn’t apply to those regular nouns, the ones with nominative plural -s!

  68. the whole question of regularity conceals a deep mystery: why do languages bother with it?
    Tax laws exist so that tax lawyers can discover loopholes in them. The purpose of IT is to fix the bugs it creates. Linguistic regularity exists so that linguists can contrast it with the irregularities.
    Regularity appears to be one of those job creation schemes. As Voltaire remarked, if no regularity existed, it would have to be invented.

  69. marie-lucie says:

    the whole question of regularity conceals a deep mystery: why do languages bother with it?
    I think that using language would be a lot more bothersome without it.
    Efficient communication needs a mix of predictability (common framework as starting point) and unpredictability (new information). Predictability resides mostly in the morphology (regular, patterned ways of creating words and adapting them in order to convey information required for the functioning of the language, such as plural inflection), unpredictability in the lexicon (vast number of words, new, old and borrowed, with meanings which can be elastic). Syntax ties words together and usually shows a mix of predictability (basic sentence framework) and unpredictability (rules of movement, deletion, merging, etc which reorganize the words of a sentence in order to affect the presentation, conveying and interpretation of information).

  70. jamessal says:

    why do languages bother with it? If people can learn 10,000 different nouns, they could learn 10,000 unique plurals for them — it’s not that much harder.
    God and everyone else knows that you know more about this than me, but that statement strikes me as strange, especially after reading Barber, who emphasizes language’s economy. Why would it not be very hard to learn 10,000 unique plurals? That languages tend to economize implies that in fact it would be hard and that’s why languages do it. That’s circular, right? But doesn’t it make sense too?

  71. I should have said that it can’t just be a memory limit (as Pullum suggested to me once). If there’s room in memory for N words, there is certainly room for 2N words.

  72. Jim: That languages tend to economize implies that in fact it [to learn 10,000 unique plurals] would be hard and that’s why languages do it [economize, i.e. don't provide 10,000 unique plurals].
    By using the blanket term “languages”, both you and John are tending to suppress the differences between languages as to plural formation and “efficiency”, thus contributing to the paradox you see but John does not. In contrast, what marie-lucie wrote avoids any paradox: “Efficient communication needs a mix of predictability (common framework as starting point) and unpredictability (new information)”. In that paragraph she speaks not of languages, but of communication – admirably Luhmannian, but way before that just (in my opinion) common-sensical. Of course neither Luhmann nor common-sense will pay the rent.
    John Emerson often referred to German noun declensions as an insuperable obstacle to his attempts to learn more German. But 80 million Germans have no particular problem with their German plurals, just as 300 million Americans have none with their English plurals. Regularity is in the mind of the beholder – that was the point of my comment above with the distorted quote of Voltaire speaking about God.
    I take John to be saying that a speaker can be accustomed to the 10,000 unique plurals in his language, because these plurals are the regularity iself – they’re not “random”, and you can even count them ! But consider: from whose point of view are there “10,000″ plurals, and what does “unique plural” mean to that observer ? 10,000 may seem nothing to an observer who speaks the language in question, but a lot to a observer whose native language is English.
    Suppose that an English speaker S hears of a language L which has 10,000 unique plurals. S thinks “cool, L must be hard to learn, I want to learn it”. S then buys a grammar of L, with a complete list of 20,000 words of which half oare the singular forms, and the other half the plurals. After memorizing a lot of the singular/plural pairs, S notices that each singular L word a1, a2 … beginning with “a” has a plural ending in “zaa”, each singular L word b1, b2 … beginning with “b” has a plural ending in “zab” – while all other words form their plurals by reduplication, i.e. the plural of “ca” is “caca”.
    S concludes that “unique plurals” is a misleading description of the plurals of L, since he has found regularities among them. S confronts a speaker T of L with this discovery, only to be told that the “regularities” S thinks he has found are superficial. It turns out that S had been studying a Monday-through-Friday grammar of L. On weekends, the plurals of L are formed in an entirely different way.
    In other words, what counts as regularity depends on what you are accustomed to count as regularity.

  73. If there’s room in memory for N words, there is certainly room for 2N words.
    Dunno about that. A weaker claim might be demonstrable: if there is room in memory for N words, there is almost room for N+a words, where a is a constant independent of N and the language in question.
    Suppose the “uniqueness” of the plural form of, say, word x in the language L consists in that form being a concatenation of x and a pseudo-random sequence of vowels tacked on to the end of x. Plural forms are recognizable as familiar words with gibberish at the end.
    The constant a would be the memory footprint of a pseudo-random number generator. If N already takes up all of memory, you would merely have to buy a small a-sized memory module to stick into the USB slot in your ear.

  74. Suppose N is large enough to accomodate a universal Turing machine S that (among other things) outputs the N singular forms of the language L. Then you wouldn’t need another N units of memory to accomodate an additional Turing machine P to output the N plurals of L. You wouldn’t need any extra memory at all, since S would just simulate P when provided with an input tape describing P.

  75. jamessal says:

    “Efficient communication needs a mix of predictability (common framework as starting point) and unpredictability (new information)”. In that paragraph she speaks not of languages, but of communication – admirably Luhmannian, but way before that just (in my opinion) common-sensical.
    Sure, but the whole point of the passage was to give a sense of linguistic evolution or, if you like, the ways languages change. It’s all well and good to pick out a simple sentence, with no jargon, and praise it for its directness or whatever else; but that jargon you find so useless you helps make the generalizations you find so boring, which in turn bear fruit for anthropologists and historical linguistics alike, in the form of reconstructed languages and what they can tell us about the people who spoke them. Did you read the part about articulatory phonetics? I’ve never seen it explained so well with so few charts and facts to memorize (I use to walk around with notecards, really) before I could proceed.
    I should have said that it can’t just be a memory limit (as Pullum suggested to me once). If there’s room in memory for N words, there is certainly room for 2N words.
    On both counts I’m still not sure why, but it’s been a rough week and I’m tired — maybe I’ll get it soon.

  76. Trond Engen says:

    John’s question of why we have patterns for plural forms is really a question of why we have grammar at all rather than purely lexical items for every possible message or emotion we might want to convey. Reductio aside, that’s really the question of why we have language at all — and how it arose.
    But for the single grammatical form: Starting from absurdum, it’s clear that there are simply too many simple messages, let alone shades of meaning, to convey in a lifetime for anyone to come across and learn the lexical item before the need for it arises. So leaning on his over-sensitive pattern-recognition tool the speaker makes an ad hoc parallel formation, and, abra cadabra, since the listener has an equally over-sensitive pattern finder, the message comes through. For every time such a pattern, or such a simple mean to convey a nuance of meaning, is used, it becomes easier to grasp for, growing in prominence until the force of analogy is balanced by the force of lexical learning for high-frequency words or other analogies formed from different patterns or by different speakers.

  77. Boy, Trond Engen’s explanation is better than whatever I would have come up with. Kudos.

  78. Trond Engen says:

    Gee, thanks. Reading it after posting I feared I had stacked too much into my sentences. And now I see that I wrote ‘… since the speaker has …’ instead of ‘listener’.

  79. Fixed!

  80. jamessal says:

    Yes, thank you, Trond — I do believe I’m seeing the light!

  81. jamessal says:

    “Did you read the part about articulatory phonetics?” I asked Grumbly. Of course not. Because that part wasn’t included in the section I pasted, long as that section was. I’ll save myself and, more so, Hat some work and not paste the part introducing articulatory phonetics now, since Stu doesn’t seem all that interested anyway.

  82. marie-lucie says:

    JC: why do languages bother with [regularity]? If people can learn 10,000 different nouns, they could learn 10,000 unique plurals for them — it’s not that much harder.
    English inflection (the modification of individual words to indicate grammatical features such as plural and others) is so simple, and the irregularities so few, that the idea of unique plurals for every one of 10,000 nouns may sound plausible.
    But take a language like Latin. Nouns are divided into 5 classes, each with 5 or 6 forms for different “cases” in the singular and almost as many for the plural. Even if there is some overlap between case forms in each of the classes, for each class of nouns you need to memorize around 8 or 10 different forms. And this is for classes, not individual nouns!
    Latin verbs are also divided into several classes. Even the most regular verbs have 6 different forms for each tense (3 “persons” for both singular and plural), and there are 6 simple tenses for the indicative mood, at least 4 for subjunctive (I forget), and yet a few more moods with fewer forms. So each single verb has over 60 forms that need to be memorized.
    Would you like each noun and verb in Latin to have a unique set of forms?
    The history of Latin and its transformation into several new languages includes simplification and regularization of these complex sets of forms (even though other complexities later arose from phonetic changes). Similarly, Old English (like German) also had several classes of nouns and verbs, even if the system was not as complex as that of Latin, especially for the verb. The simplification and regularization of English morphology was probably hastened by the intrusion of French as the dominant language for three or four centuries after 1066, but most other modern Germanic languages are also less complex than, for instance, Old High German.

  83. Grumbly, your language L clearly has regular plurals, although it’s of a type that doesn’t exist (and we don’t know why not): there seem to be no languages whose morphological rules are dictated solely by phonology. But no, I was thinking of a language whose plurals are truly random: the plural of apple is beek, the plural of bee is cati, the plural of cat is dogo, the plural of dog is eggem, the plural of egg is foxidu, and so on throughout the lexicon of nouns. Languages like this don’t exist either, but it can’t be a matter of memory constraints.
    Marie-Lucie: Latin was hard enough for me to learn without unique forms for all the morphological slots of nouns, thanks! But Latin or English, the number of inflected forms of a noun is small compared to the number of nouns, and (crucially) it doesn’t get any bigger over a speaker’s lifetime, whereas we are all constantly learning new nouns just to function in society. The very fact that languages like German and Maori exist, whose inflectional forms are mostly or completely unpredictable (though not unique), shows us that people are capable of attaching a list of arbitrary morphological facts to each item in their lexicon, and not just to a small fixed number of irregulars as in English. If we can remember which arbitrary class every noun we know is in, why can’t we use the same resources to remember arbitrary forms for each noun? And yet it seems we can’t, or at least we don’t speak languages that require us to do so.

  84. marie-lucie says:

    JC: I was thinking of a language whose plurals are truly random: the plural of apple is beek, the plural of bee is cati, the plural of cat is dogo, the plural of dog is eggem, the plural of egg is foxidu, and so on throughout the lexicon of nouns. Languages like this don’t exist …
    No, they don’t exist, because morphological markers for plural or other grammatical categories modify a word but do not normally replace it with something completely different. Except for the occasional case where an unusual series of phonological change have totally obscured the difference between two forms of the same word (as in French for “eye”: singular oeil, plural yeux, both from Latin oculus and its series of forms), singular and plural forms of the same noun will have a recognizable common stem (as in ox : oxen, box : boxes) or at least a common “skeleton” (as in foot : feet, mouse [maws] : mice [mays]). If the plural of apple is beek, there is absolutely nothing in the second form to give a hint either that it belongs with the meaning “apple” or that it is a plural of anything.
    There are rare cases in which forms of the same word are quite different from each other (as in good : better : best or go : went : gone). This is because in each case there were two (or even more) stems with very similar meanings, and for some reason each stem became used mostly for one category (such as plain versus comparative, or present versus past). We can see the beginning of this at work with person and people, the latter taking over for persons in many uses: it seems that many people is more likely to be used colloquially than many persons, and a bunch of persons is much less likely than a bunch of people. But such cases of “suppletive forms” are rare in most languages. Again, such irregularity is confined to very common words.

  85. marie-lucie says:

    jamessal: thanks for the link. I hope to read that book, I think I would agree with most of it. The link of the development of language in primitive humans with the development of language in the child has long seemed obvious to me.

  86. articulatory phonetics … Stu doesn’t seem all that interested anyway
    Thanks for sparing me that particular passage ! Like proctology, it sounds like one of those elaborate scientific approaches to something I am interested in only for pragmatic reasons.

  87. It just describes the way we make sounds by passing air from our throats past our teeth, the sounds (or phonemes: the smallest distinctive sounds in a given language) altered by the tenseness of our throats, or the placement of our tongues, or lips, etc. Allophones are even smaller units of sound, with subtler differences; each phoneme contains a group of allophones, some of which are distinctive in one language but not another. By describing these sounds in detail, these things we do with our throats and tongues as air passes through or abruptly stops, linguists are able to predict, or reconstruct with some accuracy, the ways these sounds will change or will likely have changed: through syncope, apocope, metonymy, etc. (i.e., laziness of the mouth, the dropping and garbling of phonemes). If that’s as distasteful to you as proctology, okay. I kinda like it.

  88. M.L.: No problem about the link. In fact, it was selfish, as I don’t imagine I’d get as much out of those 800 pages as you would, and you’ve got a talent for breaking it all down and passing along the juicy bits!

  89. marie-lucie says:

    I just looked up “articulatory phonetics” in Wikipedia, hoping to make a recommendation to read it. Unfortunately most of the article seems to have been written by an engineer or surgeon rather than a linguist. The main part is mostly about the mechanisms of air flow through the structures of the throat, and includes the equation for “Boyle’s law” of air pressure. Read that part if you are into laryngology, but there is little in the article that is really useful for linguistics.

  90. I just looked up “articulatory phonetics” in Wikipedia, hoping to make a recommendation to read it.
    Was it the spectacular inaccuracy of my precis that engendered this search for a recommendation? I thought I knew my fricatives from my plosives, my phonemes from my allophones, my high vowels from the low ones; but I guess maybe there are some merits to a formal education. Ah, well. At least I’ve learned that being wrong — even often — ain’t really that bad.

  91. marie-lucie says:

    The articles on phonetics and phonology are much better. The meaning of the words phoneme and allophone is well described in the Phonology article.
    The difference between the two can be illustrated by a parallelism with writing. We all know what an upper-case letter “A” looks like, but different people writing an “A” will write it a little differently, and different fonts will have slightly different shapes for this letter, as well as italic, bold, etc. Handwritten shapes can be even more different. Nevertheless, readers and writers will largely agree whether or not a certain written segment qualifies as an “A”.
    Within writing systems, the abstract letter shape, or grapheme, corresponds to a phoneme in the spoken system, for instance English /t/. The various ways (allographs) of actually writing a grapheme such as “A” correspond to spoken allophones, or actual spoken variants of a given phoneme. For instance, English speakers may be conscious of pronouncing a /t/, but the actual pronunciations of this /t/ are different in table (slightly aspirated), stable (plain or unaspirated) or (in American English) pretty (closer to a weak [d], sometimes even to an [l]). These three different sounds (which American English speakers produce effortlessly and largely unconsciously in their proper positions in words) are interpreted by the same speakers and their hearers as the single phoneme /t/.

  92. marie-lucie says:

    jamessal, there was nothing wrong with your overall description of phonetics, but not much detail in it. More detail is what I was looking for in Wikipedia, but there was (for my taste) far too much emphasis on the mechanics of voice production.

  93. I’ve learned that being wrong — even often — ain’t really that bad.
    He learnt that from me.

  94. English speakers may be conscious of pronouncing a /t/, but the actual pronunciations of this /t/ are different in table (slightly aspirated), stable (plain or unaspirated) or (in American English) pretty (closer to a weak [d], sometimes even to an [l]). These three different sounds (which American English speakers produce effortlessly and largely unconsciously in their proper positions in words) are interpreted by the same speakers and their hearers as the single phoneme /t/.
    How is that different from what I said?

    Allophones are even smaller units of sound, with subtler differences; each phoneme contains a group of allophones, some of which are distinctive in one language but not another.

    In another language the aspiration might, to go with your example, distinguish itself from the unaspirated /t/, and so that language would have a different set of phonemes from English. What am I missing?

  95. Oh, I posted my comment before seeing yours, so never mind — thank you, M.L.
    I think I stressed the mechanics of voice pronunciation because I had just read Barber on it, and found the passage so impressive; and I think I kept it short, well, for two reasons: I’m obviously not as fluent in all this as you and the other pros are, so I have to think double-hard so as not to mislead, and I also didn’t want Grumbly passing out on me.

  96. I’ve learned that being wrong — even often — ain’t really that bad.
    He learnt that from me.
    You may have a few years on me Reverend Wright, but some day, over drinks, we should try to tally our respective wrongnesses. I think I could compete.

  97. See, I was wrong just now: M.L. was saying the Wiki article had (for her taste) too much emphasis on the mechanics of voice pronunciation — not my little comment. And that’s not even the kind of wrongnesses we’ll be delving into, my friend.

  98. marie-lucie says:

    jamessal: Allophones are even smaller units of sound, with subtler differences; each phoneme contains a group of allophones, some of which are distinctive in one language but not another.
    You are right in a way, but the way you said it is rather compressed, hence (to my mind) misleading.
    Allophones are not “units of sound”, they are physical variants in the realization (actual pronunciation) of the (abstract) phonemes, which ARE the units in the sound system.
    Some phonemes have only one allophone (eg English /s/ or /f/), others can have two or more, usually depending on the position of the phoneme in the word (eg initially, finally) or its proximity ot other phonemes (eg before or after certain vowels or consonants). As individual sounds produced by the vocal tract, what is an allophone (variant) in one language can function as an independent phoneme in another. In an often-cited, aspirated plosives occur in English as allophones of plain plosives, but in many languages of India the two are separate series of plosives. An English of German speaker learning Hindi, for instance, needs to learn to pronounce a non-aspirated [t] at the beginning of some words instead of only after [s], since there could be confusion between two Hindi words identical except for the two kinds of “t”‘s. A French speaker has no trouble pronouncing unaspirated plosives, but finds it hard to pronounce English plosives at the beginning of words, since there is no aspiration in French in that position. Similarly, in English /t/ and /d/ are separate phonemes because of word pairs such as time and dime, but speakers of languages where the sounds [t] and [d] are allophones of one phoneme will pronounce the two English words identically until they learn to pronounce them differently.
    The degree of latitude in the physical realization of phonemes explains that regional or foreign speakers can have allophones that are different from those of the most accepted variety and still be perfectly understandable. For instance, many languages have a phoneme /r/ but its realization is very different in Spanish, German or French, and also in the English of various countries (England, Scotland, US, South Africa, etc). Nevertheless, as long as whatever actual pronunciation of the phoneme /r/ is different from that of every other phoneme, there is no confusion and people understand each other.
    (Note: “plosive” is the British term corresponding to North American “stop”).

  99. jamessal says:

    what is an allophone (variant) in one language can function as an independent phoneme in another
    That’s exactly what I was trying to convey with my overly compressed sentence, and your precise expansion is exactly why LH is the best blog on the interwebs. I’d had a grasp on the concept, which you’ve now tightened. And I hope I’m not the only one reading and learning. Thank you, yet again.

  100. I would say that we call a particular variant an allophone only if it makes some sort of contrast somewhere in the total system of human language. After all, no two utterances of a phoneme can be exactly alike, yet we do not speak of millions of allophones.
    My understanding of stop vs. plosive is that stop has two senses, a wider one in which it includes nasals (because the mouth is closed even though the nose is open), and a narrower one in which it does not; plosive never includes nasals. I had thought this had more to do with being a phonologist (wider sense) or not (narrower sense), and perhaps with the linguist’s age, but maybe nationality enters into it too.

  101. marie-lucie says:

    JC, you are right that we do not speak of millions of allophones. Here my comparison with writing was oversimplified in order to make the point of unit vs variant. Perhaps I should have said that “A” and “a” were allographs (depending on how generally or minutely you defined allographs). There must be a definite phonetic difference between two allophones of the same phoneme for it to be formally recognized, although in our own language we are often unconscious of such differences, while people of different linguistic backgrounds may find them obvious. Following Sapir, for a native speaker of a language, allophones have physical reality but not “psychological reality”.
    I should perhaps warn that today the most “advanced” theoretical phonology does not believe in the existence of phonemes or allophones, only of “distinctive features” (eg +stop, -nasal, etc) which are considered the basic units, although they have no individual existence but are simply descriptive terms. But I do not subscribe to this theory. I find it strange that a theory which emphasizes the role of “native speaker intuition” in syntax apparently denies it in phonology.
    Re “nasal stops”, I have encountered this terminology in some works, but it is rather rare. In most discussion of phonetics and phonology “nasal” is enough for practical purposes. For instance, (non-nasal) voiceless stops such as p, t, k are frequently aspirated, but not nasals. Otherwise rules such as the one about English aspiration would have to specify “affects stops except nasals”.

  102. He’s so competitive.

  103. More competitiver than you!

  104. For instance, (non-nasal) voiceless stops such as p, t, k are frequently aspirated, but not nasals.
    A nasal stop would be difficult to distinguish from a severe cold. Aspirated nasals would tend to interfere with breathing, but not so badly as aspirated gristle.

  105. marie-lucie says:

    Grumbly, nice puns, but the technical meanings of “stop” and “aspirated” can be misunderstood. Actually a “nasal stop” (such as “m” or “n”) stops being nasal if you have a bad cold (you say “Bob” instead of “Mom”, for instance, or “siggig” for “singing”). “Aspirated” consonants are actually “expirated” (but not “expired”).

  106. Thanks, marie-lucie. I’m never sure what “aspirated” and “stop” mean. Some day I will be inspired to have a go at phonology.
    Following Sapir, for a native speaker of a language, allophones have physical reality but not “psychological reality”.
    On the assumption that I understand what that means, I would disagree slightly – because of the role that grammatical structure appears to play. I’ll get to that in a moment.
    You know how I complained for a long time about not being able to understand French speakers unless they were “intellectuals” ? I meant by this that they pronounced more of the letters I see in written French. After listening to Arte TV for months, I picked up all kinds of phrases in their working context that don’t occur in the “intellectual” stuff I read. This made it easier for me to understand more and more of what any kind of person said.
    In another comment thread here recently, I explained this in terms of “phrasal accumulation and expectation” while listening to sentences being produced, and bulbul agreed. Above in this thread, Trond wrote what I take to be a similar view:

    Starting from absurdum, it’s clear that there are simply too many simple messages, let alone shades of meaning, to convey in a lifetime for anyone to come across and learn the lexical item before the need for it arises. So leaning on his over-sensitive pattern-recognition tool the speaker makes an ad hoc parallel formation, and, abra cadabra, since the listener has an equally over-sensitive pattern finder, the message comes through. For every time such a pattern, or such a simple mean to convey a nuance of meaning, is used, it becomes easier to grasp for, growing in prominence until the force of analogy is balanced by the force of lexical learning for high-frequency words or other analogies formed from different patterns or by different speakers.

    One could describe these “patterns” as surface grammar expectation rules. In other words, in understanding a sentence as it is being produced, a hearer does not use (learned) rules solely to parse, which by definition can take place only post-production, but also – perhaps primarily – to track and anticipate during production.
    Understanding written sentences provide more opportunity for parsing, because one can repeatedly go back and reread what was just read. There is not enough time for “deep” parsing during speech until one has learned techniques for listening to and understanding long, complicated sentences. This would involve cutting back on ongoing expectations, in favor of memorizing and post-processing. Of course a lot depends on the exact nature of the length and complexity being produced by a speaker – “long sentences” come in different flavors.
    Expectation, not parsing, seems to play a large role in both phonetic and semantic processing. Reading Dennett’s Consciousness Explained, I just came across the following paragraph:

    For instance, we hear speech in our native tongue as a sequence of distinct words separated by tiny gaps of silence. That is, we have a clear sense of boundaries between words which … do not seem to be marked by beeps or clicks, so what could the boundaries be but silent gaps of various duration – like the gaps that separate the letters and words in Morse code ? If asked in various ways by experimenters to note and assess the gaps between words, subjects have little difficulty complying. There seem to be gaps. But if one looks at the acoustic energy profile of the input signal, the regions of lowest energy (the moments closest to silence) do not line up at all well with the word boundaries. The segmentation of speech sounds is a process that imposes boundaries based on the grammatical structure of the language, not on the physical structure of the acoustic wave (Liberman and Studdert-Kennedy, 1977).

  107. The next-to-last paragraph in my comment starts with the extraneous word “Understanding”. The paragraph should start like this: “Written sentences provide …”

  108. marie-lucie says:

    Some day I will be inspired to have a go at phonology
    If you look up “Phonology” on Wikipedia, skip “Development of phonology” (which is about the history of phonological theories) and scroll down to “Overview of phonemes” (which also explains “allophones”).
    For “stop” and “aspirated”, look up “Stop consonants” and “Aspirated consonants”, both of which are short and to the point. “Consonant” also includes IPA phonetic symbols for most known consonants, and a section “Consonants and vowels” which discusses the differences.
    All of these are more readable and useful than “Articulatory phonetics” which I mentioned above and found disappointing as a linguist (although others with different tastes would probably be interested).

  109. Thanks for the reading tips, marie-lucie.
    What do you think of that Liberman/Studdert-Kennedy conclusion: “The segmentation of speech sounds is a process that imposes boundaries based on the grammatical structure of the language, not on the physical structure of the acoustic wave” ? Not much, I suspect, since otherwise you would have commented on it. The interpretation may be exaggerated, but surely there is something to it.

  110. The internet tells me that the Liberman part of the study’s authors is Alvin Liberman, father of Mark Liberman. The report is “Phonetic Perception” in the 1977 Handbook of Sensory Physiology VIII.

  111. Bathrobe says:

    Sapir wrote a famous article on the psychological reality of the phoneme. Saying that allophones have a physical reality means that they are different but won’t be heard as such because they have no “psychological reality” for the native speaker.
    The point that “The segmentation of speech sounds is a process that imposes boundaries based on the grammatical structure of the language, not on the physical structure of the acoustic wave” seems quite true to me. For instance, contrary to naive intuition, we don’t really leave gaps between words, and to do so would run the risk of rendering yourself almost unintelligible. If you have ever listened to a synthetic voice announcement that splices words together into a sequence (for example, a sequence of numbers), you will find that it sounds quite unnatural and difficult to understand. We hear speech sounds not as raw sounds, but as sounds that organised in certain patterns by language.
    (I’m just wondering, if you had a dog called Blue, it would come running if you called out “Blue!” in isolation, in a familiar Blue-calling pitch, but if you interspersed your speech with “blues”, would he react? I suspect not. However, to prove this I need to get a dog and do some experiments, which I don’t plan to do.)

  112. marie-lucie says:

    When you first hear an unknown language spoken normally (eg in a conversation), it is next to impossible to recognize any word boundaries. If speakers of this language want to communicate something to you, they will slow down and break down their sentences in order to emphasize a key word, or pronounce this word by itself, etc. This is because the sounds of the words, especially if spoken rapidly in casual speech, tend to run into each other,or at least to influence each other, since the words within word-groups (phrases) are rarely separated in normal speech (as opposed to “teacher speech”, for instance). Also, some identical sound sequences may consiste of a single word or of two or more different words, as in English apart and a part. Such sequences are easy to differentiate if spoken in isolation and very deliberately, but in the course of a conversation it would be impossible for a novice to English to identify them correctly. An English speaker of course would know which one is which, because the phrase a part and the adverb apart not only mean different things, but occur in different parts of a sentence, surrounded by different kinds of words. This is why grammatical information (knowledge of morphological and syntactic structure, whether conscious or not) is needed in order to correctly interpret such homophonous sequences.
    Sequences of words which normally occur together, such as articles and nouns, may be hard for barely literate persons, even native speakers, to be broken up. For instance, barely literate Spanish speakers often write the article and noun as a single word, especially if the noun begins with r, which is reinforced to rr initially, a sound that mostly occurs in the middle of words, so these people often write larrazón instead of la razón ‘(the) reason’.
    In French things are even worse, because words tend to be linked phonetically so as to create syllables ending in vowels if at all possible: for instance les enfants ‘the children’ sounds like [le zan fan] (an here = nasalized vowel [a]). On paper the article is written les, normally pronounced [le], but before a vowel the (very old) pronunciation of les takes over as the final s becomes the sound [z] before the initial vowel of the next word, and this [z] starts a new syllable with the vowel of en (which sounds identical to an). Most of the final consonant letters seen on paper are a leftover from Old French, the spelling of which was a faithful reflection of the pronunciation, several centuries ago. These letters are maintained in modern spelling because in many cases they must (or at least could optionally) be pronounced in certain contexts when the next word begins with a vowel: the more educated speakers are, and the more formal the occasion, the more these consonants pop up in speech, and the reverse is true of informal contexts, especially but not only with less educated speakers. That’s a major reason why people who learned French mostly from written sources find it hard to understand naturally spoken French at first. That’s also why French is not too hard for French-speaking children to learn to read, but hard for them to learn to write – you have to learn not just the individual spelling of words, but also a lot of formal grammar to write the language in the standard manner.

  113. marie-lucie says:

    Sapir wrote a famous article on the psychological reality of the phoneme. Saying that allophones have a physical reality means that they are different but won’t be heard as such because they have no “psychological reality” for the native speaker.
    Sapir was researching the Southern Paiute language with a speaker of this language, and of course writing down phonetically everything this man was saying. At some point he started to teach him how to write his language. One peculiarity of the language was that some sounds only occurred at the beginning of words, others only in the middle. For instance, there was a word Sapir wrote (I quote from memory) pava, but there were no words like papa, vava or vapa. But where Sapir had written pava the SP speaker was writing papa, even though he consistently pronounced [pava]. After several experiences of this kind, Sapir realized that in SP the sounds [p] and [v] were not distinct units, but “counted” as the same unit. Where [p] and [v] occurred within words was entirely predictable. The SP speaker might find Sapir’s pronunciation strange if he tried to say [papa] or [vava], but to him these were just unusual pronunciations of the same word for which he said [pava] and wrote papa.
    This is what Sapir meant by the psychological reality of phonemes. The SP speaker could tell that Sapir was not copying his (the speaker’s) own pronunciation properly, but the differences between [p] and [v], which were so obvious to Sapir, were meaningless to that speaker. In later terminology, they were allophones of a single phoneme /p/ in SP, even though they were different phonemes in English.
    Another example is the difference in status of the sound(s) written th, in English and in French. In English it is important to pronounce thing differently from sing, or bath from bass (the fish), as the difference is meaningful as it serves to distinguish between a number of words. In European French, saying [θ] (as in thing) instead of [s] will make you sound like a small child who has not yet learned to say [s] properly, but it won’t at all interfere with communication: all your [θ]‘s will be interpreted as [s]‘s. Even though French speakers are quite capable of pronouncing [θ] in isolation, or of imitating a child using it in their own speech, many of them find it hard to use this ability in keeping apart English words like those above: their tendency is to say sing instead of thing, because they don’t want to talk baby talk! at least at first, they don’t realize that in English [θ] is not a childish variant of [s] but a distinctive part of the English sound system: a phoneme.

  114. An English speaker of course would know which one is which, because the phrase a part and the adverb apart not only mean different things, but occur in different parts of a sentence, surrounded by different kinds of words. This is why grammatical information (knowledge of morphological and syntactic structure, whether conscious or not) is needed in order to correctly interpret such homophonous sequences. … That’s a major reason why people who learned French mostly from written sources find it hard to understand naturally spoken French at first.
    In the case of French, it took me one hell of a long time to hit on the idea that stock phrases, expectation and grammar are involved in “hearing words”, and then to take advantage of these mechanisms. It is definitely not common knowledge outside of linguistic circles. I had to reinvent the wheel all by my little self.
    No wonder people have difficulties learning a foreign language (apart from that aged neurons business) – they are never told about these facts of phonology and shown how to take advantage of them. “Language labs” are no substitute – they are like pianos placed in front of persons who want to learn how to play, but who are told “just practice over and over until you have the pieces by heart”.
    I can’t remember any such problems with German, though, nor with Spanish. Perhaps that is because I heard them all the time in everyday life, which was not the case with French – until I had the idea of letting Arte TV run 24/7 on my box. I have rarely got along well with people who require a TV muttering in the background in order to feel comfortable.

  115. m-l,
    That’s an interesting explanation of why French speakers say “s” for English “th”.
    Why is it, do you think, that instead of saying “s” and “z” for English unvoiced and voiced “th” like speakers of European French, speakers of Canadian French tend to say “t” and “d”?

  116. marie-lucie says:

    Grumbly, most textbooks intended to teach foreign languages, especially European ones for European learners (this includes English), are not written by linguists. The ideal collaboration would be between good language teachers and linguists, but that does not often happen.
    When you are among speakers of a language you want to learn, you learn it by ear, through hearing it used in actual situations, not through seeing it written on the page as in most classroom teaching situations. I learned Spanish mostly through being around Latin American friends, and reading (actual books or magazines, not textbooks) exposed me to a different style and expanded my vocabulary.
    Alan: Actually those t’s and d’s are slightly different from their t’s and d’s in French words. I think that they intend to say th but they are placing the tongue tip against the upper teeth, not in between or just at the edge of the teeth as in English. In Canada many French speakers have allophones for French /t/ and /d/ that sound almost like [ts] and [dz], before the high front vowels /i/ and /ü/ (written u), but in English they do not do so (they don’t say “tseeth”, for instance, or “dzeep”). Similarly their “th”‘s are almost but not quite [t] and [d] (eg they do not pronounce English “this” like their French “dix”, pronounced [dzIs]). But the sounds used for the English “th”‘s are similar enough to [t] and [d] to be interpreted as such by listeners unfamiliar with this variety of French.
    I confess I still have trouble with “th”‘s: if I don’t make an effort I end up sounding to others (or to myself on a tape) like I say [s] and [z] even if my tongue and teeth positions for the two “th”‘s are quite different from the ones I use for [s] and [z] in French or English.

  117. …the sounds used for the English “th”‘s are similar enough to [t] and [d] to be interpreted as such by listeners unfamiliar with this variety of French.
    But then where do those particular sounds, the ones used for English th, come from? Are they the same as the allophones for (Canadian) French /t/ and /d/ when these don’t come before high front vowels?
    Also, is the particular kind of lisping that French children do common among French Canadian children as well? If so, it would seem like the motives for using that substitution would be the same. A lisped s would still give the best approximation to an English th, no?

  118. marie-lucie says:

    Alan, I don’t really have answers to these questions, since I don’t live in a French-speaking area and only occasionally meet francophone Canadians. I can just say that to me the sounds substituted for “th” don’t sound exactly like “regular” t’s and d’s.
    It is true that “lisping” would provide the best approximation for the English sounds, but perhaps the factor of not wanting to sound like “baby talk” arises here too. It seems to me (mostly from recollection) that French children who lisp are much more obvious than anglophone adults in sticking their tongue out and pronouncing truly “interdental” sounds.

  119. One question: why do we need plural forms of nouns at all? Plurality can be indicated by inflecting other parts of speech, and languages seem to be fine with some plurals that are identical to the singular form – sheep, fish, deer in English, for example, Lehrer and Wagen in German, and in spoken French, surely, plural forms of nouns are generally indistinguishable from singular ones.
    My favourite “plural totally different from singular”, incidentally, is the Irish for woman/women, ban in the singular and mna in the plural, both regularly derived from the same Indo-European root that gave English both “queen” and “gynacology”.

  120. marie-lucie says:

    Z, there are lots of things which are not absolutely indispensable for linguistic communication, but languages have developed them over the millennia and we are mostly stuck with them, unable to even imagine that our language could deal differently with the concepts they embody.
    It is likely that the majority of prefixes and suffixes, such as those indicating plurality, were once separate words, or at least clitics, which occurred so often together with certain kinds of words that they became fused with them, and since they were not as emphasized as the main words, became reduced in their pronunciation, and sometimes eventually disappeared altogether, like plural -s in the majority of French nouns. In the case of tooth/teeth and such, the vowel alternation was not original but is due to the fact that in the Proto-Germanic ancestor the plural affix (at least for one class of words) was -i, a front vowel which affected the pronunciation of the root vowel, which underwent a change. Eventually the affix weakened to a neutral vowel -e, then disappeared altogether since the vowel difference inside the word was now sufficient to indicate whether the word was singular or plural. In German the vowel change affected a number of words, but the vowel of the suffix remained, although weakened. But plurality does not necessarily mean that nouns are marked: in one language I have studied, most nouns do not show plural (unless they indicate pairs or sets, such as hands, eyes, houses). It is mostly verbs that indicate plural: one prefix indicates that several persons are doing the same action (eg to drink; reduplication (repetition of part of the stem) indicates not the number of people doing the action but but the number of things or people affected by the action: for instance, the base form of the verb meaning to cut indicates a single cutting action, but its plural indicates repeated actions, either on the same thing (as in to slice) or on several things, regardless of how many persons perform the action.
    Plurality (of items, actors, actions, etc) seems to be important to human communication, and so is an indication of the timing or degree of completeness of actions indicated by a verb, and that is why in many languages such information needs to be specified every time that a noun or verb is used, to the extent that just about every noun or verb indicates the relevant information. For instance, in a book of history or a traditionally written novel, very few verbs do not specifically indicate that the story is about past events. One might think that titling the work “history of …” or announcing “Once upon a time …” would be enough to let the reader or listener know that the text relates past events, but instead, the several hundred verbs used in the text need to be in past form, relentlessly reminding the reader that the events in question happened at some point in the past. Why do some languages insist in such repetitive indications of plural, tense or other grammatical category, while others are quite content to use a particle here and there when this information is crucial, is a mystery.

  121. languages have developed them over the millennia and we are mostly stuck with them
    Are there any examples in English of a one-person crusade to change the language, that have actually worked? The only one I can think of is Webster’s spelling, in the US.

  122. There was Robert McCormick, a publisher of the Chicago Tribune, who started a spelling reform campaign in 1934 that ran until 1975. The newspaper printed “thru”, for example, instead of “through”.
    I don’t think one can claim that the reform “worked” – we don’t have an Academy Anglaise to adjudicate such things once and for all – but I suspect it encouraged many people to throw off the orthographic yoke, for instance advertising copy writers.

  123. marie-lucie says:

    As I have had several occasions to comment, the Académie Française does not have that much influence on the French language. As in the 17th century when the academy was set up, they meet every Thursday to work on their dictionary – so new editions are rather infrequent. Being elected to the Académie – usually after making the rounds of the members to push one’s candidacy to a vacant seat – is an honour for a writer, but I don’t think that too much concrete work is expected of the members.

  124. Stu once found out how much the Academicians were paid. I remember it wasn’t worth doing for the money. Can it have been 6 euros a day? No, it must have been more.
    I don’t think that too much concrete work is expected of the members.
    They spend a lot of time electing the new members. In the actualités, I see that there are only two candidates for fauteuil 31, but eight for fauteuil 40 – why the discrepancy, is fauteuil 40 next to the plate of brioches and the coffee? Is there a draught that comes under the door by fauteuil 31?

  125. How shaming ! Of course I meant to have written “throw off the orthographic yolk”.
    there are only two candidates for fauteuil 31, but eight for fauteuil 40
    What a bunch of lushes. They don’t need fauteuils, plain old chaises are adequate.

  126. I wonder if it has a number on the back, so you can tell if someone’s sitting in your fauteuil.

  127. I wonder if they ever play musical fauteuils.

  128. marie-lucie says:

    Apparently, the first academicians sat on chairs, except their president, who had a fauteuil. During the reign of Louis XIV a very old, invalid member asked to be allowed to use a more comfortable seat than a chair. Hearing of this, the king gave orders for enough fauteuils to be sent for everyone.
    When a fauteuil becomes vacant (through the death of the incumbent), candidates are often contacted by sitting members who ask them to apply, and they can also apply on their own. They apply to specific fauteuils because each vacant fauteuil is identified with the previous occupant. Nobody would apply to succeed a person whose work they did not like or that they had quarreled with. A few of the fauteuils are reserved for people from specific groups – I think one has to be in the military and another one in the church (this was set up centuries before the separation of church and state).
    Since candidates are elected by the members, they usually carry on a campaign by visiting members to ask for their support. After being elected, new members must have a traditional green outfit made, including a sword (designed and made especially for the member). The first official act of a new member is to deliver a speech praising his predecessor in the fauteuil. In turn, another member (chosen by some ancient rule) “welcomes” the new member with a speech which is often quite critical. These two speeches are usually published in the major newspapers.
    This sort of conservatism is more often associated with Britain rather than France.

  129. marie-lucie says:

    to succeed a person : this doesn’t seem right, but I can’t think of an alternative. Please help!

  130. I can’t think of an alternative either, but that’s because it seems right. In the present context you could even say “succeed a person to the chair”. Besides that, there is “succeed a person to a position” or “succeed a person as …”.
    It those sound slightly strange, that may be because they are phrased abstractly, like dictionary entries. In actual use the person and position are concrete: “succeed M. Vitroux to the fauteuil”, “succeed Mr. Smith to the presidency”, “succeed Mr. Smith as president”

  131. I just learned that “a lush” doesn’t mean what I thought it did, but rather “a heavy drinker”. I suppose I was trying to make a noun out of “luxuriate”. There’s also the French luxure meaning péché de la chair, here in the sense of fishing for fauteuils.

  132. This sort of conservatism is more often associated with Britain rather than France.
    Yes. Maybe that’s why I find it amusing.
    new members must have a traditional green outfit made, including a sword (designed and made especially for the member)
    But the pen is mightier than the sword.

  133. Trond Engen says:

    Yes. Maybe that’s why I find it amusing.
    In a sworded way.

  134. My wife was talking about the Royal Botanic Gardens in London this morning, and I’m a bit deaf, so she was shouting, and she said “For Kew!” At fist I couldn’t think why she’d suddenly got so upset with me.

  135. (I should add that she hardly ever swears, though I do.)

  136. But the pen is mightier than the sword.
    Under the rule of men entirely great. But where do you find those?
    “The pen is mightier than the sword, but the sword speaks louder and stronger at any given moment.” —Sir Roger Fenwick, founder of the Duchy of Grand Fenwick

  137. Under the rule of men entirely great,
    the pen is mightier than the sword – yet since
    The sword is longer and more stiff, there will
    Be those who much prefer it to the quill.

  138. This gets even worse if you omit the space after pen.

  139. Under the rule of She Who Is Obeyed,
    The scabbard grips the drooping hen-pecked blade.

  140. if you omit the space after pen
    Wikipedia: The motto appears in the school room illustration on page 168 of the first edition of Mark Twain’s The Adventures of Tom Sawyer (1876). The words “pen” and “is” are suspiciously close together leading some scholars to speculate that the illustrator, True Williams, deliberately chose the narrow spacing as a subtle obscene prank. (See Ensor, Allison R. 1989. “Mightier Than the Sword” An Undetected Obscenity in the First Edition of TOM SAWYER. 27. Mark Twain Journal. p. 25.)
    Qui plume a, guerre a. (To hold a pen is to be at war). Nigel Voltaire.

  141. He who writes, fights.
    He who cites, rights fights.
    Etc.

  142. Who is Nigel Voltaire ? Is it this guy ?

  143. a sword (designed and made especially for the member)
    And what is this sword supposed to do to said member? Whatever it is, it sounds painful, maybe even lethal. Those French kings expected a lot of their subjects….

  144. 1694-1778. Nigel was just a sort of nickname. Everyone called him Nigel – everyone except Rousseau, who called him Nigella or Big Niggy.

  145. Bathrobe: I read that to mean that the sword, though very short, was also exceedingly sharp.

  146. marie-lucie says:

    According to Wikipedia under “habit vert”, the Académie was originally part of the “Maison du Roi”, the kind’s household, and the members were therefore considered as noblemen and allowed to wear the sword. This made them all equal, as there were both nobles and non-nobles among the members.
    The “habit vert” is a more recent introduction, the suit has gone through several designs and colours, always with heavily embroidered collar and lapels.

Speak Your Mind

*