Parrot communication

From: Wesley R. Elsberry (welsberr@inia.cls.org)
Date: Tue Mar 14 2000 - 11:42:37 EST

  • Next message: Tedd Hadley: "Re: Poll Finds That Support Is Strong for Teaching 2 Origin Theories"

    Stephen Jones wrote:
    >On Thu, 9 Mar 2000 18:01:16 -0600 (CST), Wesley R. Elsberry wrote:

    WRE>[...]

    I'm going to add in some numbers for referencing what Stephen
    says and claims in the following.

     SJ>... Alex, an African gray parrot, is unlike any other
    >animal: he can talk. When he says, "come here!" he really
    >wants his owner to come here. That is remarkable to the
    >scientist studying him, Irene Pepperberg of the University
    >of Arizona. ...

    Stephen snipped a bit from what I quoted. I will
    restore it, because it is relevant.

    SEJ>[...] Alex, an African gray parrot, is unlike any other
    SEJ>animal: he can talk. When he says, "come here!" he really
    SEJ>wants his owner to come here. That is remarkable to the
    SEJ>scientist studying him, Irene Pepperberg of the University
    SEJ>of Arizona. She says Alex understands that words have
    SEJ>meaning and he does not just mimic random sounds he has
    SEJ>been taught. "These birds have the emotional and social
    SEJ>skills of about a 2-and-a-half year-old, 3- year-old
    SEJ>child, says Pepperberg. "Their intellectual skills are
    SEJ>more like a 5 or 6 year old in some cases."
    SEJ>Differentiating Objects' Characteristics Alex can identify
    SEJ>about 50 different objects, can name seven colors, and
    SEJ>knows numbers up to six. "You can ask him what color, what
    SEJ>shape, what material," says Pepperberg, "and he knows what
    SEJ>set of answers belong in those categories."... Once again,
    SEJ>the line blurs between humans and animals. The one
    SEJ>remaining distinction, in the end, may be that humans are
    SEJ>better at things, but it is still surprising what a bird
    SEJ>with a walnut-sized brain can learn. ...

     SEJ>[This is a big
    >problem for those who claim that chimps and gorillas can
    >talk.[1] If the claim is that chimps can really use sign
    >language because they are closest to humans[2], then what is
    >the explanation for a *parrot* who talk as well, if not
    >better?[3] I saw a parrot sing "Happy Birthday" in an
    >opera-singer voice at the Singapore bird park but I no one
    >claimed that it knew what it was singing.[4] Parrots are just
    >very clever mimics[5] and human beings are very good at
    >training them[6] and reading into their pets' behaviour their
    >own human feelings[7]. Maybe this exposes as an
    >anthropomorphic delusion the whole field of talking apes?[8]]

     WE>Stephen says that a parrot that can communicate is a problem
    >for ape communication studies. His conclusion does not follow
    >from his premises, but this is not unusual in these little
    >bits of commentary. It is obvious that other support than
    >close phylogenetic ties have to be made for the reasons that a
    >parrot can handle interspecies communication showing concept
    >understanding.

    SEJ>It might help if Wesley first of all try to understand what
    SEJ>his opponent was *really* saying before he attempts to refute
    SEJ>him.

    It usually helps if my correspondent can clearly say what my
    correspondent means. I can only try to decipher what my
    correspondent *actually* writes, not what my correspondent
    might try to reconstrue it as after the fact.

    Notice that Stephen specifically ignores what was reported about
    Alex and his understanding of concepts when Stephen made his
    universal claim in [5]. Perhaps Stephen needs to get some
    practice in reading for understanding himself.

    SEJ>I said nothing about "that a parrot that can *communicate* is
    SEJ>a problem for ape communication studies" (my emphasis). Of
    SEJ>course parrots can "communicate". Parrots, dogs, cats, and all
    SEJ>animals AFAIK can "communicate" with members of their own
    SEJ>species by cries, barks, meows, etc, and maybe even with
    SEJ>members of other species?

    SEJ>The real question I raised "is whether chimps and gorillas"
    SEJ>and parrots "can *talk*" (my emphasis). An Associate Professor
    SEJ>in Linguistics on the other List I was on (I cannot give his
    SEJ>name so don't ask), said in response to my same post on that
    SEJ>List:

    SEJ>"...she would have to prove a lot to show that the
    SEJ>communication of Alex the Parrot is the same kind of
    SEJ>communication in human language which uses combinatory
    SEJ>elements at the morphological and syntactical levels which not
    SEJ>only refer to things in the real world, but also things in the
    SEJ>"unreal" world as well as symbols that refer to other
    SEJ>symbols."

    It might be a good idea for Stephen and his correspondent to
    see just how far Pepperberg has gone toward these desiderata.
    I don't think that Pepperberg is claiming that African grey
    parrots have the capabilities of fully nuanced language
    production and understanding, but she has impressively
    documented several components of the cognitive skills needed
    for language understanding and use at some level.

    Here's a popular account:

    [Quote]

    Then Pepperberg holds up three spools of different sizes and
    colors."Which is smaller?" Show Alex a paper triangle and ask,
    "What shape?"and he'll say, "Three-corner." Show him five
    Popsicle sticks dyed red and ask, "What color?" and he'll say,
    "Rose." Then ask, "How many?" and he'll say, "Five." He is
    clearly responding to the question itself, as well as to the
    objects. He understands "different" and "same" and can answer
    questions about relationships: Show him a blue-dyed cork and
    blue key and ask, "What's the same?" he will answer, "Color."
    Show him two identical squares of rawhide and ask, "What's
    different?" and he will say, "None." Substitute a pentagon for
    one square, and he will answer,"Shape."

     [...]

    The initial aim was to teach Alex to use words for objects,
    something no bird had been proven to do before. Next the
    focus moved to categories of color and shape, to numbers, to
    concepts such as similarity and difference. At every stage
    Alex was subjected to rigorous tests. The results had to be
    above question---beyond any suggestion the bird was receiving
    cues from the researchers. Tests were administered by students
    who had not taken part in the training. Pepperberg kept score
    of Alex's answers but sat with her back turned, unable to see
    the objects being presented. Each response had to be clear
    enough to be understood. Alex would get no hints, no leniency.

    But as the tests became more complex, Alex continued to score
    around 80 percent accuracy in his answers, far above what
    would have been possible by chance alone. Carefully
    documenting the parrot's progress, Pepperberg published one
    scientific paper after another.

    [End Quote -
    <http://www.cages.org/research/pepperberg/subisalx.html>]

    And another...

    [Quote]

    The results obtained with these techniques are
    striking.(1990b) (1990c) Alex has mastered tasks once thought
    to be beyond the capacity of all but humans or certain
    nonhuman primates. Not only can he produce and comprehend
    English labels, but he also understands concepts of category,
    "same/different", absence, quantity, and size.

    Alex has learned over 40 object labels: paper, key, nut, wood,
    wheat, truck, hide (rawhide chips), peg wood (clothes pins),
    grain, cork, corn, walnut, block, box, showah (shower),
    banana, pasta, gym, cracker, scraper (nail file), popcorn,
    chain, kiwi, shoulder, rock (a lava stone beak conditioner),
    carrot, gravel, cup, citrus, back, chair, chalk, water, nail,
    grape, grate, treat, cherry, wool, green bean, and banerry
    (apple). We have tentative evidence for labels such as bread
    and jacks. He has functional use of "no", phrases such as
    "come here", "I want X", and "Wanna go Y" where X and Y are
    appropriate labels for objects or locations.

    He has also acquired attribute labels. He can identify 7
    colors, rose (red), blue, green, yellow, orange, grey, and
    purple. He labels 5 different shapes as 2-, 3-, 4-, 5-, or
    6-cornered objects. He uses "two", "three", "four", "five",
    and "sih" (six) to distinguish quantities of objects,
    including groups of novel items, heterogeneous collections,
    and sets in which objects are randomly arrayed. He combines
    attribute and object labels to identify proficiently, request,
    refuse, categorize, and quantify over 100 different objects,
    including those that vary from training exemplars. His
    accuracy averages approximately 80% on tests of these
    abilities.

    Alex comprehends a limited concept of "category". He has
    learned not only that "green", for example, is one instance of
    the category "color", but also that, for a particularly
    colored and shaped object, "green" and "3-corner" represent
    two different categories of its markable attributes. Thus he
    categorizes such objects with respect to either attribute
    based on our vocal query of "What color?" or "What shape?"
    Because the same exemplar can be the subject of shape and
    color queries at different times, Alex must be able to change
    his basis for classification. Such ability to "reclassify" is
    thought to indicate the presence of "abstract aptitude".

    Alex has also learned abstract concepts of "same",
    "different", and to respond vocally to the presence or absence
    of information about these concepts. Thus when shown two
    identical objects or ones that vary with respect to some or
    all of the attributes of color, shape, and material, Alex
    responds with the appropriate category label as to which
    attribute is "same" or "different" for any combination; if,
    however, nothing is same or different, he replies "none". He
    responds accurately to objects, colors, shapes, and materials
    not used in training, including those for which he has no
    labels. Furthermore, he is indeed responding to the specific
    questions, and not merely on the basis of his training and the
    physical attributes of the objects: His accuracy was above
    chance levels on questions such as "What's same?" for a green
    wooden triangle and a blue wooden triangle. If he were
    ignoring the question and responding on the basis of his prior
    training, he would have determined, and responded with the
    label for the one anomalous attribute (in this case,
    "color"). Instead, he responded appropriately [i.e., "shape",
    though "mah-mah" (matter) would also have been correct].

    We have further data supporting his comprehension skills. Alex
    was able to view different collections of 7 exemplars (each
    collection chosen from among 100 objects of various
    combinations of shapes, colors, and materials), be asked 1 of
    4 possible vocal questions, each of which requested a
    different type of information (e.g., "What color is
    object-X?") about a single object in the collection, and reply
    vocally to each question. A correct response indicated that he
    understood all the parts of the question and used these parts
    to guide the search for the one object in the collection that
    provided the requested information. Alex responded with an
    accuracy of 81.3%. We also have evidence for comprehension of
    his own use of labels: If a trainer incorrectly responds to
    Alex's requests (e.g., substitutes an unrequested item), he
    generally responds (approx. 75% of the time) by saying "No"
    and repeating the initial request. Finally, a soon-to-be
    published manuscript demonstrates that he comprehends a
    relative concept: He responds correctly to questions - asking
    which of two objects is the bigger or smaller, and responds
    "none" if they are of equal size.

    These findings have at least two important
    implications. First, remember that my training paradigm could
    be designed only by integrating results obtained both from
    birds and humans, in the field and in the laboratory. Second,
    the data indicate the existence of complex cognitive
    capacities in a species whose brain organization is
    considerably different from that of terrestrial and aquatic
    mammals. That his abilities are comparable to those of
    mammals, and that such capacities have been found in so
    unexpected a source, suggests that we must keep our minds open
    to the capacities of others. At least that is what I reflect
    upon when Alex greets me each day with "Hiyo...come here!"

    [End Quote -
    <http://www.cages.org/research/pepperberg/harvard.html>]

    A bibliography through 1997 is available at
    <http://eebweb.arizona.edu/Faculty/Bios/pepperberg.html>

    A good page with a fair number of links is at
    <http://www.cages.org/research/pepperberg/index.html>.

    Perhaps Stephen and his correspondent can catch up to this
    research sometime soon. It's not like Pepperberg's work with
    Alex is breaking news or anything. My visit to her lab was in
    1993, and she had already contributed several seminal papers
    to the peer-reviewed literature by that time.

    Nor are parrots and non-human primates uniquely identified as
    having some of the basic cognitive capabilities for language
    use. Lou Herman at the Kewalo Basin Marine Mammal Research
    Lab in Honolulu has performed many experiments showing concept
    understanding on the part of bottlenose dolphins using commands
    with a structured syntax. Ron Schusterman at the Long Marine
    Research Lab at UC Santa Cruz has established the ability of
    sea lions (Zalophus californianus) to form equivalence classes.

    There has been a lot of work done that addresses the issues
    that Stephen raises. Why Stephen should act as if these
    concerns had *not* been addressed or even considered
    previously is a mystery.

     WE>Why this might impact ape communication
    >studies is something Stephen does not explicate.

    SEJ>I actually asked a *question*: "If the claim is that chimps
    SEJ>can really use sign language because they are closest to
    SEJ>humans, then what is the explanation for a *parrot* who talk
    SEJ>as well, if not better?"

    SEJ>Maybe Wesley can have a try at answering it?

    Stephen did ask a question, but Stephen did not *limit* himself
    to inquiry, as his statement referenced in [8] clearly shows
    him dabbling in unsupported speculation.

    I had a go at explaining why Stephen's question was ill-posed.
    It is obvious that Stephen is using "talk" in multiple senses
    which are not clearly compatible with each other. My
    discussion split out components. Interspecies communication
    is necessary, but not sufficient. This does not necessarily
    imply concept understanding on the part of the non-human, but
    it may open up avenues of inquiry into such concept
    understanding. Stephen's question assumes that only one
    underlying explanation can be adduced for all instances of
    interspecies communication with concomitant concept
    understanding on the part of the non-human. That is an error
    on Stephen's part. The reasons why non-human primates may
    have both communicative and cognitive skills required for
    language use (if they do have them) need not be the same
    reasons that parrots may be able to communicate and have
    understanding of the concepts communicated. Thus, Stephen's
    question is of the "Have you stopped beating your wife?"
    category.

    Nevertheless, I did set about discussing how we can go
    about trying to put together an explanation of parrot
    cummunication and concept understanding capabilities, as
    Stephen requested. As such, I did "have a try" at
    answering the parts of Stephen's question that could be
    split out and made sense of. The nonsensical parts,
    like the assumption of one singular explanation fitting
    all examples, were left to their own devices.

     WE>Why should a parrot be able to handle concept understanding
    >and interspecies communication? The interspecies
    >communication part is made possible via mimicry, but mimicry
    >alone does not explain the whole set of phenomena. For
    >explaining that, more needs to be known about African grey
    >parrot ecology and social interactions. One of Pepperberg's
    >graduate students, Spencer Lynn, was recently in the field in
    >Africa with a research team looking into the ecology and
    >behavior of African grey parrots in the wild.

    SEJ>I have no problem with "interspecies communication".

    That's good.

    SEJ>My problem was the claim that animals can *talk*.

    I included "concept understanding" in my discussion. If we
    are being skeptical all around, I wonder if *Stephen* can
    *talk*. Someone capable of *talking* might consistently and
    coherently state their position. Perhaps Stephen could say
    what, precisely, delimits his usage of "talk" such that we can
    move on. Does "talk" strictly mean vocal utterances as the
    medium, or is ASL OK by Stephen?

    SEJ>My cat meows for his food and I can understand what he is
    SEJ>communicating. My wife claims he has different meows for
    SEJ>different things, but I can't pick it. A parrot might actually
    SEJ>learn that the mimicked words "polly wants a cracker" caused
    SEJ>food to be offered and then says those words in order to get
    SEJ>food. But that does not mean that the parrot knows what the
    SEJ>words actually mean. If it doesn't, then it is not talking as
    SEJ>humans talk to each other.

    Which shows that "concept understanding" actually does address
    Stephen's concerns. If "concept understanding" is necessary
    but not sufficient to meet Stephen's connotation for "talk",
    we will need the other conditions specified before going on.
    Stephen conveniently ignores the fact that I phrased my reply
    as a conjunction of "concept understanding and interspecies
    communication", not just "interspecies communication". Why
    would Stephen do that, I wonder?

     WE>I've met Pepperberg and Alex, and spent some time at the lab
    >in Tucson watching the research going on. Irene Pepperberg is
    >a careful researcher whose methods include extensive
    >documentation and analysis of all interactions when Alex is at
    >his "work" station (the back of a lab chair). Alex scores
    >about 80% correct on various types of questions put to him.
    >This level of performance is far above chance. I knew Spencer
    >Lynn from his time at Texas A&M University at Galveston, where
    >he took part in the large-scale assessment of sperm whale
    >abundance and distribution in the northwestern region of the
    >Gulf of Mexico (the GulfCet project, funded through Minerals
    >Management Service). These people are very capable
    >researchers, and I find Stephen's flippant assessment of their
    >activities to be close to libelous.

    SEJ>First, since when is questioning a scientific claim "libellous"?

    When it is done in a way that calls the basic competence of
    the researcher into question in almost complete ignorance
    of the facts of the matter. In other words, like what Stephen
    does in his original statements. Stephen attributes
    anthropomorphism in [8]; all that is left is *who* Stephen
    was attributing it to. By the evidence of [5], [6], and
    [7] it seems clear that even if Pepperberg is not specifically
    named, she certainly falls into the category of humans who
    interact with parrots. The "this" of [8] then links the
    claim of anthropomorphism to his previous commentary.

    SEJ>Second, if Wesley took the trouble to actually read my
    SEJ>comments above, to out what I actually said (as opposed to
    SEJ>what he wants me to have said), he would discover that in my
    SEJ>comments I said nothing about "Pepperberg and Alex", at all!

    Stephen's comments were made in the context of discussion of
    the web article on Pepperberg and Alex. The very first word
    of Stephen's comment [1] is "This". What might the referent
    for "This[1]" be if not the research discussed in the web
    article that Stephen had just quoted? That article, in case
    Stephen has forgotten it already, was about Pepperberg and
    Alex. Stephen's comment [3] only makes sense in this
    discussion as a referent to Pepperberg and Alex. Stephen's
    comment [8] would be a complete non sequitur if not in
    reference to Pepperberg and Alex. What does "this" in
    Stephen's statement refer to if not Pepperberg and Alex? If
    Stephen were referring to non-research parrots kept only as
    pets, why would that have any bearing upon what the web
    article putatively implied about ape communication? It looks
    like Stephen either *was* commenting upon Pepperberg and Alex,
    or Stephen was engaging in yet another deceptively packaged
    non sequitur. I don't see a third alternative in there, but
    Stephen is welcome to have a go at inventing one
    after-the-fact. That seems to be one of Stephen's better
    talents.

    I think my reading of the words that Stephen actually wrote
    has been sufficiently careful. Stephen, though, doesn't seem
    to be able to practice what he preaches, given his focus upon
    one element of a conjunction I used to the exclusion of the
    other element.

     WE>If Stephen thinks that he can form a novel cogent and valid
    >critique of Pepperberg's work, I encourage him to pick up some
    >of her publications and try it out. Otherwise, Stephen is
    >(as usual) blowing smoke.

    SEJ>See above. I was not actually making a "critique of
    SEJ>Pepperberg's work".

    As long as we are in nit-pick mode, I didn't claim that Stephen
    had made a critique, but rather challenged Stephen to do so
    in the future.

    However, one can read Stephen's comments as such a critique,
    and I think it most natural that they would be read as such a
    critique. Stephen's comments were in discussion of a web
    article on Pepperberg and Alex. Stephen's comments make a
    universal claim about parrots in [5], another claim about
    human training of parrots in [6], and makes a conclusion in
    [7]. Given that Stephen did not qualify these premises or his
    conclusion, one is left with the impression that Stephen
    believes that Pepperberg's work is an instance of
    anthropomorphic delusion. If Stephen does not actually
    believe this, it would be good for him to explicitly disclaim
    it.

    Stephen's uninformed and prejudiced commentary made in the
    context of discussion of Pepperberg and Alex certainly is
    "blowing smoke" on his part. To correct this, Stephen would
    have to learn something about the actual research and be able
    to make informed commentary upon it. So far, this appears to
    be well beyond Stephen's capabilities. A novel cogent and
    valid critique of Pepperberg's research would demonstrate that
    appearances were deceiving in this case. (Such a criticism
    that was not novel could simply be due to "mimicry" on
    Stephen's part, and thus would not count.) Until then,
    though, appearances don't appear to be deceiving.

    SEJ>My comments were directed at the "field of talking apes".

    I assume, then, that Stephen's usage of "talk" is broad enough
    to include ASL-using non-human primates.

    As mentioned above, Stephen's comments were not *solely*
    directed to "talking ape" studies, but rather made claims
    about parrots, language understanding, and the people who
    study the two (see [5], [6], [7], and [8]). Stephen also
    clearly references parrots as a subject in [1]. [3], and [4].
    It is *ludicrous* for Stephen to try to say that his "comments
    were directed at the "field of talking apes"" as if they were
    *exclusively* so.

    SEJ>Any "smoke" that Wesley is experiencing is probably from
    SEJ>his own spinning wheels!

    Such wit. I guess that puts me in my place. Stephen squirms
    a lot here, but doesn't seem to address my criticism of his
    comments. The research that Stephen doesn't think possible
    already is underway and has been for years. The evidence
    showing concept understanding on the part of non-humans
    continues to grow, aided by experimental studies such as those
    performed by Irene Pepperberg. It is plain that there can be
    multiple ways in which elements of language and concept
    understanding can be instantiated in animals, and there is no
    need to try to make the same reasons apply to all instances,
    as Stephen's original question required.

    Some other resources concerning Irene Pepperberg and Alex...

    <http://www.wnet.org/nature/parrots/html/intro.html>
    <http://www.physics.helsinki.fi/whale/intersp/who.html#Alex_the_Great>
    <http://www.geocities.com/Heartland/Prairie/9935/alex.html>
    <http://www.azstarnet.com/nonprofit/alexfoundation/>
    <http://www.suite101.com/article.cfm/world_languages/22888>
    <http://www.2think.org/parrots.shtml>

    SEJ>"When speaking here of Darwinism, I shall speak always of
    SEJ>today's theory that is Darwin's own theory of natural
    SEJ>selection supported by the Mendelian theory of heredity, by
    SEJ>the theory of the mutation and recombination of genes in a
    SEJ>gene pool, and by the decoded genetic code. This is an
    SEJ>immensely impressive and powerful theory. The claim that it
    SEJ>completely explains evolution is of course a bold claim, and
    SEJ>very far from being established." (Popper K., "Natural
    SEJ>Selection and the Emergence of Mind," Dialectica, Vol. 32,
    SEJ>Nos. 3-4, 1978, pp.339-355, pp.343-344)

    Well, that particular bold claim is false, given that Popper
    does not include genetic drift as a contributor to an
    explanation of evolution. The framers of the modern
    synthesis, though, were one up on Popper and included it.
    Given that the modern synthesis was put together in the
    1930's and 40's, it seems odd that a scholar like Popper
    would fail to have a grasp of it by 1978. Of course, that
    is the same article where Popper finally figured out that
    natural selection is not just a metaphysical research
    programme, again several years after the light should have
    dawned.

    Wesley



    This archive was generated by hypermail 2b29 : Tue Mar 14 2000 - 11:16:50 EST