A clever parrot learns to combine phonemes (not)

« previous post | next post »

No matter how hard I try to locate the world's most stupid animal communication story, they keep outflanking me. I am always left behind. An even stupider one always comes along. All I can say as of this morning is that I never thought I would see a story as stupid as this in a respected news source, and right now I cannot imagine how it could be surpassed (though within a few weeks I suppose it probably will be). The Economist has published (10/25/08:103) a review of a new book called Alex & Me in which Dr Irene Pepperberg tells the story of her scientific life with Alex the grey parrot (see here and here for a couple of Alex's earlier appearances on Language Log Classic). The Economist has already shown a certain affection for Alex's story: it devoted its obituary of the week to Alex when he died in 2007. The review calls the new book "a memoir of two unusual scientific careers, one of them pursued — not exactly by choice — by a bird." Now, I should make it clear that I do not have the book. If this merited scholarly investigation I would of course obtain it; but given what I know so far, I am deeply reluctant to part with $23.95 to get hold of a trade book for sentimental parrot fanciers (the subtitle is: "How a Scientist and a Parrot Uncovered a Hidden World of Animal Intelligence — And Formed a Deep Bond in the Process"). So I will simply tell you about the stunningly stupid part of the review, and leave it to you to determine, if you care to, whether the review misrepresents the book on this point. But I warn you, especially if you know a little elementary articulatory phonetics, that this one will boggle your mind. Are you prepared to face the rest of the day with a boggled mind? Then read on.

The review is anonymous, like all Economist reviews, and after what I will say below, the reviewer will doubtless come to see this as quite a blessing. What the reviewer reports, relying on the book, is that Alex the parrot didn't just chat with his keeper and keep her entertained, and know how to name fifty objects and the numbers from 1 to 6, and combine words to make up expressions (Alex is said to have named cake "yummy bread", though why this is counted as a creative act rather than a failure to learn the simple word cake is not clear from the review); the thing is, Alex also "seemed to combine phonemes to construct new words". This is a true first: creative lexical word formation from phonological segments, in a bird. An example is cited:

Lacking lips, he could not pronounce the letter "p", so his term for an apple was "banerry" (apparently mixing "banana" and "cherry").

Well, you probably caught it. The phonemes (systematically distinct speech sound units) that humans produce by bringing together the upper and lower lips, known as bilabial consonants, include [p] as in pop, [b] as in bob, and [m] as in mom. Say these words while looking in a mirror and you will see that your lips come together. And when you say "banerry" or "banana", of course, it happens just the same. Apparently the reviewer did not have the intelligence to reflect on why failing to learn one word with a [p] in it should have something to do with not having lips, when pronouncing [b] was apparently no problem for the self-same bird. I stared at the page, almost unable to believe what an asinine thing had been printed there. (I have no idea whether the reviewer's remark has a basis in the text of the book. Quite possibly it doesn't.)

The sheer dumbness of the reviewer's remark gets even worse if you reflect a little more. Parrots don't have upper front teeth either, or the bony gum-covered area behind them that is known as the alveolar ridge, or a velum (the soft membrane at the back of the mouth) that can be lowered to open the velic port between pharynx and nasal cavity, so by the logic of the above quote they should also be unable to pronounce [n], a nasal consonant produced by lowering the velum and making a closure in the oral cavity by pressing the tongue tip against the alveolar ridge. In fact parrots have just about none of the articulatory apparatus that humans have, which should mean they are utterly without the ability to mimic any human speech, right?

Of course not. They mimic the sound of human utterances brilliantly. But they do it in a totally different way, simulating the acoustic effect (not the articulatory production) by means of an organ known as the syrinx, which birds have and humans don't. This clever flute-like organ works for mimicking all kinds of sound without using anything like the organs or movements humans make do with. Parrots don't need the anatomical attributes of the human oro-pharyngeal tract any more than an iPod or a radio does. One of the staggering things about the quoted statement is that the reviewer didn't appreciate that obvious point. I suppose it's just one more measure of how little knowledge educated people have about phonetics. After a hundred years of really detailed work (three thousand years if you want to take things back to the age of the brilliant phoneticians and phonologists of ancient India) the study of the general nature of speech sounds and how they are produced is still almost unknown to the general public. And the typical undergraduate career in a university includes not one single course on elementary aspects of the science of language.

There's more to be said about the review, and doubtless, far more than I'm ever going to say about the book. But taking [b] to be a suitable substitute for [p] in a creature with no lips looked to me like definitely the dumbest piece of science writing about animal communication this month. (It's October 29; there's only two more days for some stupid fake pet communication trick story to beat it.)

[P.S.: I get a lot of flak for being skeptical about whether people's pets understand everything they say and name lots of objects and communicate with them better than a human child could etc. etc. But, mindful of the spirit of democracy that animates our great blogosphere, I have opened comments on this post. So below you will probably soon be able to read a long series of boring grumbles and sentimental special pleading by gullible bird-lovers who hate me. Enjoy.]



56 Comments

  1. Nik Berry said,

    October 29, 2008 @ 8:33 am

    So, seriously, this reviewer has never heard of a parrot saying "Pretty Polly" or "Pieces of Eight"?

    [Added by GKP: Oh, thank goodness, a smart commenter is first up. Yes, I had intended to mention this, because of course it did occur to me, but it got lost as the fury took hold of me and I wrote more and more. (I'm better now, thank you.) "Pretty Polly": the classic stereotype of a parrot's verbalizations. Both words with a [p]. How could the reviewer possibly not have thought of that? The answer to how the parrot says such things is of course that it can indeed mimic the acoustic effect of a voiceless bilabial stop before a vowel using its syrinx. No lips are needed if the job at hand is merely producing a rough acoustic simulation of human speech sounds. Oh, and look, already another commenter below has noticed that "Polly" begins with [p].]

  2. Orange said,

    October 29, 2008 @ 8:41 am

    "Bolly wants a cracker."

  3. Elizabeth McCullough said,

    October 29, 2008 @ 9:10 am

    "I suppose it's just one more measure of how little knowledge educated people have about phonetics."

    I would say it's more a case of a little knowledge being a dangerous thing. As you say, anyone with lips, educated or not, can readily discern that "p" and "b" are pronounced bilabially. Either the book is stupid, or the reviewer was substituting a tiny amount of half-understood phonetics for a great deal of native intelligence.

  4. Roger, FCD said,

    October 29, 2008 @ 9:20 am

    You left out the hypothesis that a parrot's vocal tract has trouble with imitating a 'p' sound directly after an 'ah' sound.

  5. Michael Tinkler said,

    October 29, 2008 @ 9:21 am

    You know, there IS the used market – that way neither the author nor the publisher is rewarded!

    I'm seeing the book starting at $10!

  6. Nik Berry said,

    October 29, 2008 @ 9:26 am

    @Roger:

    "Who's a pretty boy then?"

  7. outeast said,

    October 29, 2008 @ 9:48 am

    I suspect that the 'p' thing is a reviewer error rather than a replication of a foolish statement in the book. But whatever.

    I hesitate to say or ask anything for fear of the ire I may incur, but… how is a creation such as 'yummy bread' not a 'creative act'? My son is learning to talk, and when he lacks an appropriate word for something he quite often 'names' it by combining known words in this kind of way, often to amusing effect; even my wife (whose native language is not English) does something similar (she once greatly amused me by coining 'big spoon for the garden' for 'shovel').

    A coining such as the one cited sounds to me like a real (if simple) use of language; and that would surely hold even if it was accompanied by a simple failure to learn the word 'cake'. Why the dismissiveness? I really don't have a horse in this race one way or another, but I'd like to hear a bit of a better explanation for why this is not creative language use.

    PPS I didn't know that stuff about the syrinx. Very interesting!

  8. Mark P said,

    October 29, 2008 @ 10:03 am

    I haven't read the book and don't intend to, but absent information to the contrary, I would guess that the reviewer is, indeed, parroting something he read in the book.

  9. K said,

    October 29, 2008 @ 10:04 am

    Forget new or used–do you want it taking up your precious shelf space? Get it from the library!

    Dead horse I guess, but my aunt & uncle's bird can say, "Where ya goin'?" Try saying "where" without lips.

  10. Jens Fiederer said,

    October 29, 2008 @ 10:16 am

    I was guessing a spirit of masochism rather than a spirit of democracy, but it hasn't been too bad thus far.

    I DO have issue with the used market, though – even buying it on the used market provides economic benefit to the author/publisher, although indirectly.

    The ability to resell a book can be factored into the original purchase price – if people believe they can recoup a large portion of a book by reselling it (instead of treasuring it forever or using it for landfill), they are willing to pay more up front, at least in theory.

  11. Mark F. said,

    October 29, 2008 @ 10:31 am

    It always seemed to me that a certain amount of imagination was required to hear some of the stop consonants, including /b/ and /p/, in parrot vocalizations. Even if true, this of course does nothing to change Geoff's underlying point.

    On the broader issue of animal language and cognition, I think way too much attention is paid to the question of whether some animal or another has crossed some intellectual Rubicon into territory once thought to be ours alone. The difference between a difference in kind and a difference in degree is only a difference in degree.

  12. Neal Goldfarb said,

    October 29, 2008 @ 10:50 am

    @outeast: . . . how is a creation such as 'yummy bread' not a 'creative act'?

    More to the point, if Alex really uttered (squawked?) yummy bread in reference to cake, and not just randomly, can't that be seen an example of predication?

    I'm not saying that it's necessarily predicational; it might just be indicational — i.e., two separate comments about the same thing (roughly comparable to That was yummy. That was bread.)? But it doesn't seem outrageous to me to suggest that it's an example of predication. Indeed, Geoff's colleague James Hurford argues in his recent book The Origins of Meaning that animals are capable of forming simple propositions.

    Of course, this isn't to say that Alex possessed language. But Geoff's apparent dismissal of yummy bread as trivial does seem surprising.

    [GKP: My good friend Jim Hurford and I argue enthusiastically about many things, but not about whether there are animals that can form simple propositions in their heads. It's quite reasonable to think, as Jim does, that some animals can come to an awareness of certain truths. What I insist on is that no non-human animal has ever had the nifty idea of expressing a proposition overtly with (acid test!) the sole aim of knowingly causing some other creature to be better informed. Private appreciation of what is the case, yes; deliberate informing of other creatures concerning what is the case, no. And notice, I'm not going to take a sequence of hand signs meaning "BANANA BANANA SWEET GIMME BANANA FOOD GIMME" as a suitable example, because it's transparently just a desperate attempt to obtain a banana. Given the ingenuity of animals in obtaining food for free, we really need to focus on non-food subject matter. The point of expressing the proposition has to be to update some other organism's information state, not to obtain some food (or other desirable thing) for oneself. That is going to involve having a full awareness that some other organisms do have mental lives and contentful information states, as well as ability to control a language that is up to the task of encoding propositions. Chimps don't seem to have the former, and attempts to show that they have the latter have failed. Dogs seem to have a little bit of the former but no sign at all of the latter. And parrots? Just large feathered sound reproduction devices, I suspect; not (sorry, Dr Pepperberg) mind-aware, language-capable, intelligent beings. I know nothing about the circumstances surrounding the coining of the alleged phrase "yummy bread", of course. But I suspect that attributive modification of nouns by adjectives has nothing to do with it, and that the emotional bond between Dr Pepperberg and her beloved animal would have been an insuperable obstacle for any linguist trying to explore the question. People who have trained up an animal they have come to love never seem eager to let an independent linguist or psychologist into the picture. See the anecdotes in my "Monkeys saying things again — not".]

  13. JBL said,

    October 29, 2008 @ 12:10 pm

    To paraphrase B.F. Skinner: the real problem is not whether animals think but whether men do.

  14. Timothy Martin said,

    October 29, 2008 @ 12:23 pm

    Has it even been shown that parrots are capable of learning words as symbols and not mere signs? (i.e. that a parrot can learn that "cake" refers to an actual object in the world, and isn't just a sequence of sounds it's supposed to make when someone shows it a cake?) Because if parrots don't have this ability, I don't see how a parrot could decide on its own to describe cake using "yummy" and "bread" (and putting them in the right order, no less!)

    [GKP: Timothy's error here, I believe, lies in imagining that the parrot is doing anything that could be correctly called "describing". All the interpretation here is being done by humans. We know nothing much about the parrot except that on at least one occasion Dr Pepperberg thought it made a squawky noise rather like "bread" just after it had made a squawky noise rather like "yummy". Leaping from that to talk of description is exactly the sort of insane over-generous imputation that we, the most generous animals on the planet, are so amazingly inclined toward.]

  15. Ryan Denzer-King said,

    October 29, 2008 @ 12:29 pm

    Geoff's last comment raises an interesting point for me. (Just for reference, let me make clear that while clearly animals have communication, I do not accept the proposition that animals have language.) I've worked with several developmentally disabled children who I would wager have essentially no concept that other organisms have inner mental lives and contentful information states, especially those children towards the severe end of the autism spectrum. They treated other people essentially as animated objects, where speaking was in essentially every case merely self-expression (private appreciation) or imperative (a la "banana banana"). Of course none of these children had any problem with language; all were completely fluent in everyday speech. So my question is, how does this bear on your point? Am I simply wrong, and I simply didn't understand their worldview? Or is this awareness of the inner mental lives of others something inherent in the human animal, something which doesn't necessarily need any conscious awareness of the Other? Because I would swear on the holy book of any religion that several of the children I've worked with have absolutely no concept of the Other, nor any concept of themselves as an object with an outside perceivable by others with contentful information states.

    [GKP: Deep question. It is this sort of dissociation of linguistic behavior from theory of mind, or from other mental capabilities, that makes so many linguists believe that a predisposition toward language, conferring elaborate syntactic abilities, is innate in humans. See The Mind of a Savant by Neil Smith and Ianthi Tsimpli, for example. I tend to think linguists often overstate the case here, and imagine that all the work to demonstrate the existence of this easily triggered innate mental ability to do syntactic analysis has already been done. I think the issue of what is innate (other than a sort of general thinking ability) is wide open. And I suspect that the blindness of autistic humans to other humans' information states may be exaggerated sometimes. But if it could be thoroughly demonstrated that a profoundly autistic human having no more awareness of other minds than a chimpanzee could acquire fully normal language use? That would be very interesting, and perhaps profoundly important.]

  16. N said,

    October 29, 2008 @ 1:06 pm

    I remember this exact thing from a National Geographic article on animal cognition:


    "Under Pepperberg's patient tutelage, Alex learned how to use his vocal tract to imitate almost one hundred English words, including the sounds for all of these foods, although he calls an apple a "banerry."

    "Apples taste a little bit like bananas to him, and they look a little bit like cherries, so Alex made up that word for them," Pepperberg said. "

    [http://ngm.nationalgeographic.com/print/2008/03/animal-minds/virginia-morell-text]

    I don't buy it. Phonetic similarity is irrelevant if the explanation isn't correct. I'd be happy to try to hook a parrot up to an ERP task or some sort of reinforced lexical decision task to see what they can actually produce.

    [GKP: The question of whether the reviewer added the absurd detail about lacking lips is taken up by another commenter below. By the way, I note that Dr Pepperberg's claim about apples tasting "a little bit like bananas to him" is utterly impossible to substantiate, given the private character of taste sensations — what philosophers of mind would call qualia. Alex never reported any taste similarity judgments. Dr Pepperberg is not speaking as a scientist here. I think she is making stuff up. Though there is another comment below on that topic too, believe it or not.]

  17. Christian DiCanio said,

    October 29, 2008 @ 1:22 pm

    I'm currently teaching a Gen-Ed course on Animal and Human Communication. As a linguist with a little background (a little I stress) in ethology, I've found that the discussion people have around animal communication is often polemic. Where I agree that it has not been shown that animals are capable of the full complexity of human language or even its basic complexity, there are certain abilities that border on what we might call phonology or syntax.

    For instance, birds with large song repertoires often use discrete "song packets" to build individual songs. These packets contain a series of elements, such as two short chirps at 5 kHz and one trill at a lower frequency, which are then put together to create a song. This is arguably only the case for birds which have many songs in their repertoire, such as nightingales (with up to 200 songs!). However, it seems similar to phonology insofar as discrete units are combined to create larger objects. In the case of birdsong, these songs as a whole do not have very nuanced semantics. It's essentially always to attract a female with on'e's vocal talent (male birds in most oscines do the singing, not females).

    Insofar as syntax is concerned, there is little evidence in the ethological literature that any species has it. Certain chimpanzee vocalizations tend to co-occur, which has led some researchers to posit a "basic syntax" (Crockford & Boesch, Call Combinations in Wild Chimpanzees, Behavior 142:397-421) (although the "two-word" stage is seldom considered evidence of syntax in the language acquisition literature).

    While it is fair to approach the question of "language" in animals with caution, I also believe that it would be folly to dismiss research by informed ethologists who are critically examining the issue.

    If you want any references, please ask.

  18. Kutsuwamushi said,

    October 29, 2008 @ 1:27 pm

    Wouldn't the importance of "yummy bread" depend on whether or not Alex was taught that as a vocabulary word?

    I haven't read anything about Alex, but it seems reasonable to me that researchers might attempt to create new vocabulary out of existing vocabulary so that Alex wouldn't have to learn new sound patterns.

  19. Nik Berry said,

    October 29, 2008 @ 1:34 pm

    It's not unlikely that bananas tasted a little like apples to him. The main flavour compounds in apples are 2-methylbutyl acetate, butyl acetate, and hexyl acetate. The main one in bananas is isoamyl acetate.

    (But ask someone to smell some isoamyl acetate, and they won't say "banana" – they'll say "pear".)

  20. S Hawkins said,

    October 29, 2008 @ 2:06 pm

    I haven't read the book, but a New Yorker article last May suggests that Pepperberg herself is more careful about her claims:

    "Pepperberg will never really know what it is like to be a parrot, and she is careful not to claim that Alex learned a language. She calls what he learned a “two-way communication system,” and prefers the more conservative term “labels,” rather than “words,” to describe his vocabulary."

    Obviously the general public has not paid any attention to this distinction in their rush to anthropomorphize the parrot. It is lamentable that so many people see communication and language as essentially the same thing.

  21. Marivic said,

    October 29, 2008 @ 2:13 pm

    In response to Geoff's response to N:
    I read the National Geographic article when it came out and remember that author talking about lips too. No mention of the syrinx though. Here's what I found in the link N posted:

    "Shape," Alex said. His voice had the digitized sound of a cartoon character. Since parrots lack lips (another reason it was difficult for Alex to pronounce some sounds, such as ba), the words seemed to come from the air around him, as if a ventriloquist were speaking. But the words—and what can only be called the thoughts—were entirely his.

  22. Sili said,

    October 29, 2008 @ 3:05 pm

    (three thousand years if you want to take things back to the age of the brilliant phoneticians and phonologists of ancient India)

    I'd rather like to hear more about those fellas.

    There was a stupid article in the Politiken the other day about the impending death of the Danish language. I'm happy to report that (with a few unhappy examples of 'language rage') the commenters largely called bollocks of the tirade.

    (If anyone cares enough, I'll have a go at translating it.)

  23. James C. said,

    October 29, 2008 @ 3:41 pm

    S Hawkins: “It is lamentable that so many people see communication and language as essentially the same thing.”

    Lamentable in some senses, however it tells us something very deeply insightful about human thought. Humans are so focused on language as their primary communication system that they tend to interpret *all* forms of communication as language. It’s as if a lemur was interpreting speech sounds and body movements in terms of smell. To humans, language and communication *are* largely the same thing, to the point that hucksters make lots of money “teaching” humans to use communication systems other than language, like body “language”, eye movement “language”, etc.

    In Alex the Parrot’s case, humans are at an even greater disadvantage than usual in ethologic observation and interpretation of animal communication. Why? Because Alex uses sound as the communication medium rather than body postures and movements, smells, or the like. Humans are inevitably bound to interpret these vocalizations as language, because that’s all we normally think of when we think of audio communication. It takes a careful scientific viewpoint to disassociate vocalization from language. I reserve my criticism for the ethologists being unscientific rather than the lay public, and in this situation it seems that Dr. Pepperberg has done very well in maintaining a neutral interpretation.

  24. Christy Mason said,

    October 29, 2008 @ 4:28 pm

    Another interesting subject – I have only personally known one bird that could speak and it just sang commercial jingles – it's owner left it in the room with the television on all the time. As for my own three cats and dogs, I pretty much think that if they could communicate with me, their statements would be limited to "Feed me," "Play with me," "Let me out" and "Stop bothering me." I've seen the Alex story in the press over the years and what I recall is that the claim was that Alex had learned labels for things, and that he put the labels together, e.g. identifying small red ball and large blue bell correctly. It was never clear to me if he was given food rewards for each correct identification. I'm pretty sure that if my dog could vocalize, he would learn to recite the Pledge of Allegiance for a piece of cheese. I've always felt that for pets, food = love = food. It doesn't hurt my feelings that they love me best when I feed them.

  25. Christian DiCanio said,

    October 29, 2008 @ 4:50 pm

    But to be clear, I find the Economist reviewer's comments to be completely stupid.

  26. Mark P said,

    October 29, 2008 @ 4:51 pm

    There was a Far Side cartoon that showed the results of a dog bark translator. Every bark was translated as, "Hey!"

    Here is one link:
    http://nachofoto.com/photo-of-Farside-Dog-Translator-E42279258e9b1

  27. Lance said,

    October 29, 2008 @ 5:06 pm

    Perhaps that Economist reviewer had been confused by reading Samuel R. Delany? Alas, Google Books gives me only this snipped view of a critique of one of his works:

    If Delany was teasing us he teased too well. Brass can't pronounce a p because "the mouth, distended through cosmetisurgically implanted fangs, could not deal with a plosive labial unless it was voiced"; how could voicing [to b] make all the difference? … More imagination about oral impediments was needed.

    Alas, from the snippet, I can't even tell what work is being discussed, though Brass is a character in Babel-17.

  28. mud and flame said,

    October 29, 2008 @ 5:09 pm

    Here's what Pepperberg says about it in The Alex Studies: Cognitive and Communicative Abilities of Grey Parrots, as revealed by Amazon's "look inside the book" feature:

    …We thus made an exception to our rule against training food labels, and in fall 1984 began training "apple." At that time Alex already used the labels "banana," "cherry," and "grape." During formal sessions, he began to produce a /p/. At the end of the season for fresh apples, he refused these fruits, and his vocalizations remained at this level. We thus removed apples from training and the laboratory. Apples were reintroduced in the spring, were eaten, and /p/ reappeared in the first training session. During the second week of training, however, Alex looked at the fruit, said "banerry . . . I want banerry," and snatched a bite. He not only persistently identified the fruit as "banerry" in subsequent sessions, but also slowed production and sharpened his elocution ("ban-err-eeee"), much as trainers do when teaching a new label (Pepperberg 1990c).

    Most likely, the reviewer misunderstood a version of this account in the newer book. (My experience is that most book reviews contain at least one factual inaccuracy about their subjects.)

  29. Elizabeth said,

    October 29, 2008 @ 5:22 pm

    The bilabials in "yummy bread" weren't a problem then?

  30. Bobbie said,

    October 29, 2008 @ 5:34 pm

    Sorry, but my mind keeps wandering to Chicken Lips, as in the lyrics sung by Bruce Springsteen:
    Oh, chicken lips and lizard hips
    And alligator eyes
    Monkey legs and buzzard eggs
    And salamander thighs
    Rabbit ears and camel rears
    And tasty toenail pies
    Stir 'em all together
    And it's mama's soup surprise"

  31. Karen said,

    October 29, 2008 @ 6:24 pm

    Nor, apparently, was the P in "grape". Weird.

  32. Ray Girvan said,

    October 29, 2008 @ 8:25 pm

    K > Try saying "where" without lips

    Not difficult; you just pronounce the "w" using only the palate (i.e. replace the voiced labial-velar approximant "w" with the similar-sounding velar approximant "ɰ").

    Lance > Brass is a character in Babel-17

    Yep; that's where it's from. Brass, the ship's pilot, has been surgically modified to have lion-like features (fangs, claws, mane) and can't say "p". Delany's not consistent, however, in one conversation (p86) he says "This thing must be big as a damn battleshi' …. Ca'tain Wong" where only a few lines earlier he's said "There haven't been any hijacks on transport ships for for seventy years".

  33. teucer said,

    October 29, 2008 @ 9:08 pm

    Geoff, you've said a couple times that you don't believe animals have ever formed a declarative sentence or asked a question. Personally, I'm not sure I agree – but it's certainly true if you use the word "question" narrowly.

    My mother's dog, when she was still little enough not to have learned which new things she found were chew toys and which she'd get scolded for chewing, used to do something I can only interpret as soliciting information. She would pick the thing up very gently in her mouth, then look up at my mom, who would either praise her or gently say "no-no." The response determined whether her dog would subsequently chew the item with gusto or leave it alone forever.

    Now, I believe that to be communication (and the fact that some sort of communication exists in animals is well-established) and it's clearly intended to solicit information. But it isn't forming an interrogative sentence; it's not a use of signs (be they characteristic barks or movement or anything) that have specific meanings independently being put together in a way that is a question.

    So to state that it isn't language is trivially true – but at the same time, I think it implies that she is communicating in a way that requests a communicated reply containing information, rather than really being on par with the repeated "BANANA" sign example. It could even suggest an awareness on the dog's part of my mom as a being with a mind capable of storing information (and a different set of information from that stored in the dog's brain), but I wouldn't be prepared to advance that particular claim about dog intelligence without more conclusive evidence ideally of a less anecdotal nature.

    [GKP: Listen carefully, teucer; I vill say zis only once. This is not Communication Log. This is not Information Log. This is Language Log.]

  34. James said,

    October 29, 2008 @ 10:26 pm

    Speaking of grumbles… Dude :) Substanitive analysis without ridicule works just as well.

    [GKP: James has a problem here. He has implicitly accused a person in a very powerful position — a senior writer for the great Language Log corporation — of being an intellectually bankrupt grumbler who substitutes ridicule for substantive analysis. But he has misspelled "substantive" in his hasty comment. And Language Log does not have a facility for commenters to edit their comments after submission. So my moral quandary is, should I grumble at the misspelling, and ridicule him? Or should I show some intellectual responsibility and substantively analyze it? Let me see… Don't rush me. I'm thinking, I'm thinking…]

  35. the other Mark P said,

    October 29, 2008 @ 11:23 pm

    [quote]The main flavour compounds in apples are 2-methylbutyl acetate, butyl acetate, and hexyl acetate. The main one in bananas is isoamyl acetate.[/quote]

    Esters are the main taste component of loads of fruit. An apparent similarity in one part of the name ("acetate") means nothing, since that is commonplace in fruit esters. Ethyl acetate is used as nail polish remover, and it is very similar in structure to butyl acetate.

    Quite apart from that, who is to say that parrots stress the ester component of taste as much as humans? Perhaps they "taste" the different sugars, for all we know.

  36. Timothy Martin said,

    October 30, 2008 @ 12:16 am

    @teucer: Imagine I am staying over the house of a friend who is currently sleeping, and I am moving around the house trying to make as little noise as possible. I come to a door that I need to open. I move it gingerly at first, trying to see if it's going to creak or not. If it does creak, I continue moving it slowly until it's open enough for me to enter. If it doesn't creak, I move it more quickly, confident that it's not a creaky door and I needn't be concerned that I'm going to wake my friend.

    …Was I soliciting information from the door? Maybe in a sense, but I don't think it proves any linguistic ability, and I don't think it's any different from the example you gave of your mother's dog. The dog apparently knew that bad things sometimes happened when it chewed on new objects, just as I know that a door can creak and wake someone. Your mother's reaction was a sign that it would be alright or not alright for the dog to proceed, just as the sound made by the door would be to me a sign that it would be necessary to proceed more carefully or not.

    [GKP: Timothy, you are beautiful. You hear the music. This is what I should have said to teucer if only I had your ability to do substanitive analysis. Lovely example. Come back often. We need you.]

  37. Alex said,

    October 30, 2008 @ 10:56 am

    "'(three thousand years if you want to take things back to the age of the brilliant phoneticians and phonologists of ancient India)'
    I'd rather like to hear more about those fellas."

    I think he was referring to Panini (~2.5 thousand years ago), but there may be others. He is the most well known.

  38. David Hilbert said,

    October 30, 2008 @ 11:55 am

    I was struck by the following comment from Professor Pullum:

    By the way, I note that Dr Pepperberg's claim about apples tasting "a little bit like bananas to him" is utterly impossible to substantiate, given the private character of taste sensations — what philosophers of mind would call qualia. Alex never reported any taste similarity judgments. Dr Pepperberg is not speaking as a scientist here. I think she is making stuff up.

    I have no idea what evidence Pepperberg had or didn't have for her claim about how similarity in taste was perceived by Alex. What I do have some idea about is the study of perceptual similarity in respect of color and it is not usually considered necessary, or even desirable, to elicit explicit similarity judgment when studying similarity in perceived color in human beings and non-human animals. There are a variety of techniques that are used to infer perceived similarity from a variety of data. There is a long tradition of inferring similarity from discriminability and with somewhat fewer assumptions inferring similarity from choice behavior. In human beings, at least, these techniques can be compared to the results of eliciting explicit similarity judgments and they work reasonably well with the kinds of caveats that are found in any area of science. Pepperberg may be a fraud for all I know but not because claims about similarity concern mysterious qualia that are inaccessible to science. To give one example from the literature here is a paper about color similarity for honeybees:

    Backhaus, W., R. Menzel, et al. (1987). Multidimensional scaling of color similarity in bees. Biological Cybernetics 56(5): 293-304.

    I think this is real science and hasty references to qualia are not helpful here.

  39. Alex said,

    October 30, 2008 @ 11:09 am

    I must admit I find the tone in this post a bit off-putting. Regarding the reviewer Pullum writes,

    "I suppose it's just one more measure of how little knowledge educated people have about phonetics."

    I agree, but does this then merit phrases or statements like

    "sheer dumbness of the reviewer's remark"

    "the reviewer did not have the intelligence to reflect on…"?

    Is it really a matter of intelligence? Why not just say "the sheer wrongness of the reviewer's remark"? To claim that it is just dumb is to confuse fact and opinion. You may be able to cure bad analysis, but not unintelligence, so why criticize something you cannot change?

    Pullum by the way reminds me of Colin McGinn!

    [GKP: My tone is off-putting? I really must beef it up a bit! I don't want to be merely off-putting. I want to frighten people with my fury. Frighten them into paying a little attention to the possibility that we might apply careful thought to questions about natural language sometimes. Listen, Alex: sometimes I do argument, and sometimes just abuse. Today I am not doing argument. This isn't argument. This is abuse. (And hey, at least give me credit for not making up any stupid jokey remarks about whether you're related to the other Alex!)]

  40. Cephi said,

    October 30, 2008 @ 12:48 pm

    @Alex: what "dumbness" accomplishes, that "wrongness" does not, is a more or less explicit indication that the reviewer has been culpably intellectually lazy. A writer can be wrong without being blameworthy; but to be dumb (in the relevant sense) is to be a bad writer. When writers write badly, we should hold them accountable for it.

  41. Marc A. Pelletier said,

    October 30, 2008 @ 1:23 pm

    Hey, look! My favorite topic!

    I agree fully with prof. Pullum here; the whole claim is farcical and, well, just plain dumb. And I'd take the claims of "yummy bread" with doubt, myself.

    An interesting question some of the comments raise, however, is "*Is* there a definition of 'language' that (a) distinguish it from 'communication' (either as strictly distinct or as a superset, (b) can be objectively assessed and (c) doesn't pressupose humans are involved?"

    I mean, there is no question that animal communication takes place, that human communication takes place, and that a great deal of human communication takes place through language; but if we define language as "certain types of human communication" then the statement that no animal other than humans use language is true but vacuous.

    I guess an interesting question would be that, were we to find an alien or some other unfamiliar creature, what would convince a linguist that they use a language in the absence of familiarity or quantitative method of evaluating intelligence (not that "intelligence" is any easier to define)?

    — MAP

  42. jamessal said,

    October 30, 2008 @ 1:35 pm

    I guess an interesting question would be that, were we to find an alien or some other unfamiliar creature, what would convince a linguist that they use a language in the absence of familiarity or quantitative method of evaluating intelligence (not that "intelligence" is any easier to define)?

    There are Charles Hockett's 13 Design Features of Language: http://en.wikipedia.org/wiki/Charles_F._Hockett#Hockett.27s_13_Design_Features_of_Language

  43. Marc A. Pelletier said,

    October 30, 2008 @ 1:59 pm

    @jamessal: I don't think they apply very well; they're good hypotheses of why human language works the way it does and how it distinguishes itself from other form of terran communication (for lack of a better generic term) but they wouldn't be applicable "tests" to determine if some other species had a language.

    I haven't read the 1959 paper, mind you, and there might be further subtlety to Hockett's reasoning that I am unaware of that addresses that question.

  44. Marc A. Pelletier said,

    October 30, 2008 @ 2:08 pm

    @jamessal: (cont'd) In fact, perusal of the Wikipedia article you've pointed out leaves me entirely unconvinced that Hockett seriously attempted to find defining points of language so much as justify (circularily) the assertion that humans use human language because that's the language humans use the way they use it.

    Some of the "design features" beg the question, some are "merely" necessary for *any* communication to take place, and the observation that only human language have all 13 of his proposed features is entirely vacuous given that they are a list of features _of human language_. I can trivially list dozens of communicative features of other species that humans to not share – this speaks nothing about linguistic ability.

  45. jamessal said,

    October 30, 2008 @ 2:08 pm

    they wouldn't be applicable "tests" to determine if some other species had a language.

    No, I suppose they wouldn't be. I didn't realize how specific your question was. Thanks for clarifying.

  46. jamessal said,

    October 30, 2008 @ 2:22 pm

    Marc A Pelletier: I first came across Hockett's approach through John McWhorter's Teaching Company lecture series on language and then in David Crystal's Encyclopedia of Language. Not being a linguist, I always just took it on faith (those guys are usually pretty trustworthy) that this is what defines language — though I did also have a creeping feeling that it was, as you say, a little circular. Then again, I'm not quite ready to call any of the particular items question-begging or arbitrary, mostly because I just don't know, but also because Crystal's Encylopedia entry seems to differ from the Wiki. Maybe the Wiki oversimplifies and misleads.

  47. Marc A. Pelletier said,

    October 30, 2008 @ 2:33 pm

    @jamessal: That's also quite possible; the problem with the simplifications that generalist encyclopedia must sometime impose on topics (and Wikipedia is no more immune to that effect) is that sometimes critical subtleties can be ommited because they would otherwise require too much prose to place in context.

    The reason why I'm a little more convinced that the proposed reasoning is at least mostly circular isn't in the fine print, but in the nature of the list itself: it does a good job at showing human use language because it lists caracteristics known to exist in human language(!) If there was a list of "good indicators" that a form of communication is linguistic in nature, then humans would probably match many, maybe most, of them… but all of them? That's a good indication that the criteria were observed in humans and declared "defining features" ex post facto.

  48. Marc A. Pelletier said,

    October 30, 2008 @ 2:37 pm

    @jamessal: … and indeed, looking at some of these factors, I'm not even convinced they are necessary for linguistic communication at all – "Broadcast transmission and directional reception", for instance, appears to be entirely unrelated to the linguistic nature of communication unless Hockett is arguing that speaking over the phone isn't linguistic in nature!

  49. Alex said,

    October 30, 2008 @ 3:59 pm

    @cephi who said: "what 'dumbness' accomplishes, that 'wrongness' does not, is a more or less explicit indication that the reviewer has been culpably intellectually lazy."

    For simplicity let's take one of Pullum's statements: "Apparently the reviewer did not have the intelligence to reflect on…"

    If only the reviewer weren't lazy he would have more intelligence! I agree with you we should hold authors "accountable" in a sense (though I don't like that term because it reminds me of politics). But that does not mean we need to speak in terms of stupidity or intelligence. Those are beside the point. They are just non-starters for inciting the change looked for, which I take it is to get people to know more about language and linguistics. Mark Liberman is a great example on Language Log (see e.g. "Sexual Pseudoscience on CNN").

    I think the reviewer would be insulted if he read Pullum's post. I could be wrong, but anyway I think the reviewer *should* feel excited to learn that there is more to that what he was saying. That's not accomplished by saying he doesn't have the requisite intelligence to reflect on his own analysis.

  50. Geoff Pullum said,

    October 30, 2008 @ 4:57 pm

    Listen, I really object to my tone being called (in Alex's earlier comment) "off-putting". I don't want to be just off-putting. I want to scare people with my fury. And maybe frighten them into paying a little attention to the possibility that it might be worthwhile to apply careful thought to questions about natural language sometimes. I have obviously been much too polite. Listen, Alex: sometimes I do argument, and sometimes just abuse. Today I am not doing argument. Now, as it happens, I have been able to find out, through secret channels that I cannot divulge (Language Log has its sources), who the reviewer was. And indeed, he is a brilliant man with philosophical credentials and many intellectual accomplishments. He couldn't conceivably be described as unintelligent. Would he be insulted if he read my post? I certainly hope so; I don't want to believe I'm losing my touch. The point is that this isn't argument; this is abuse. What part of "I'm mad as hell and I need to rant" do you not understand, Alex? Do I have to be the calm analytical professor of general linguistics all the flaming time? I'm taking an evening off from that. So I rant about claims concerning the phonetic abilities of dead parrots. Call it a hobby. It's less violent than going out and shooting and butchering a moose, isn't it?

  51. jackofhearts29 said,

    October 30, 2008 @ 5:33 pm

    I'll take a shot (as a layman and an avid reader of this blog)

    What's wrong with defining "language" as "a fuzzy subset of communication, consisting of a repertory of signifiers that are acquired from external sources during development, rather than from innate behavior patterns"?

    This would leave out autonomic gesture, laughter, etc. as well as animal calls, but would include signing, written and spoken language, and conventional/culturally specific gestures.

    In other words, "a communication system consisting of representative and conventional signs, transmitted culturally."

    Of course this definition also circles back on itself.. but that is to be expected, since we are naming a set whose purpose is specifically to encompass human language as we know it…

  52. Alex said,

    October 30, 2008 @ 6:45 pm

    @Geoff Pullum who said: "What part of 'I'm mad as hell and I need to rant' do you not understand, Alex?"

    I certainly understand that. I think we are (or could be) in agreement: given the goal of persuasion, maybe insult and facts about intelligence don't work; given the goal of ranting, insult and facts about intelligence work great. Maybe readers like ranting because they get riled up themselves. Nothing like a Two Minutes Hate. It will do no good to have a sloppy debate about meta-goals.

    But anyway, the price you pay for raging against false statements is making some of your own. (Did his stupidity really prevent him from reflecting on what he was saying? By your post, definitely not.)

  53. dr pepper said,

    October 30, 2008 @ 7:01 pm

    @Nik Berry

    I was told back in college chemistry that banana essence was iso acetyl toulene. It certainly seemed that way when i poured some out.

  54. Stonecutter said,

    October 31, 2008 @ 11:37 pm

    So then, to be clear, The Economist is often wrong or just sometimes?

  55. Sault said,

    November 1, 2008 @ 5:25 am

    @dr pepper
    Do you know what you meant by iso acetyl toluene? I can't work out what you would mean by that name (iso usually implies a non-straight chain isomer, but both acetyl and toluene cannot have anything but a straight chain), and a google search turns up nothing. Also, Ihad a quick look for what the taste in banana is and the sources seemed to support Nik Berry's isoamylacetate.

  56. Taylor B said,

    September 2, 2012 @ 9:07 pm

    @Geoff "the study of the general nature of speech sounds and how they are produced is still almost unknown to the general public." Yeah, wow, it is amazing how shallow introspection into ones language is when its done by people who aren't linguistically aware. I had a friend in a graduate math program, who was pretty damn smart when it came to problem solving and always seeing the general picture of things, but one day when we were talking about linguistics he said "lol linguistics is not a science". Its exactly these types of articles that give impressions like my friend had.

    Some people think linguistics is an outgrowth of literature and poetry, and these people are not aware of the mathematical formalisms that underlie linguistics, even if they are possibly familiar with mathematical formalism and its role in other sciences. No one ever thinks that the science of language also has its mathematical models. Why will people readily accept that electricity is handled by the physicist, software handled by the computer scientist, management handled by the businessman, medicine handled by the doctors, but when a grammar debate arises, no one will admit their ignorance, refrain from speculation, and think to ask the linguist? All I can say to this is that in general the average human being has a very high tolerance for ambiguity and is habitually far too willing to make and accept oversimplifications.

    Anyways, I am not exactly sure why you are so offended by this and your response so acerbic? Who cares if the public is linguistically illiterate or miseducated? I know businesses and religion exploits the scientific illiteracy of people (religion being a true threat to humanity), but linguistic illiteracy can only be capitalized on by the prescriptivist, and so what? Its a dumb mistake yes, but this hardly seems worth discussion, especially personal insults. Anything anyone says about anything to do with language and who has not studied linguistics will most likely be incorrect, and it will always seem increasingly absurd with further scrutiny.

    This type of aggressive reaction from anyone in the linguistic community is going to turn people off to discussions of language or linguistics, even though your intention is completely the opposite. Someone should have written a letter, and The Economist published an addendum in a later issue and have the review rewritten for the archives.

RSS feed for comments on this post