Oh Koko!

We'll be sharing much more, in the coming months, about why Anne and I are a wee bit obsessed with Koko the gorilla. She's one of the stars of a project we've got going--a project that will likely diminish our chances of ever meeting the lovely and extraordinary creature.

It's because we've been stalking Koko that Anne and I were perplexed, but not surprised, by this recent piece in The Atlantic:

Click here for the article.

Our eyebrows arched just reading the headline.

It leads a reader to anticipate a conversation with Koko the gorilla, but instead it is mostly an interview with Koko's main human, Penny Patterson. So, I guess the headline is describing what that interview is about--a conversation about conversing. There's no other way to make sense of it. Yet, it's still problematic because gorillas don't converse. Not even Koko.

The piece begins with an anecdote about Koko describing herself with the "queen" sign. Morin, the article's author, follows it with this quote of explanation from Patterson:

Koko understands that she’s special because of all the attention she's had from professors, and caregivers, and the media.

And this is where, to my mind, any interviewer actually fascinated with Koko's mind asks: This is amazing...how could Koko understand that she receives more human attention than most other gorillas on the planet? And how does she make the connection between this concept, this existence of hers, and that of a queen? When you taught her "queen," what was the definition? 

Unfortunately, this isn't the direction the piece goes. That nugget was a foray into her backstory about her sign language training and her home in California.  And just because the written product of Morin's interview doesn't dive into questions like those italicized up there, it doesn't mean it didn't. But if he did go there, why on earth didn't that gold make it into the piece?

More to the point, why have I never read an interview about Koko that asks penetrating questions about the stories that Koko's humans tell about her mind? Is that nuts and bolts stuff merely uninteresting fodder for readers or is it off-limits to writers who are permitted access? There's an answer at the end of this blog post.

But first, I have more, seemingly infinite, questions... How does Koko's language ability, and let's add Kanzi's too, compare to a highly-trained border collie like Chaser?



Is it categorically different, indicating an actually different kind of cognition?
You might be thinking about this video of Kanzi and saying, yes, yes it's different:



But how is it different? Kanzi is just pairing up two things, which is not really all that different from Chaser getting one toy at a time. It might appear to be more extraordinary than it is because in pairing the two objects for which Kanzi knows the English words, he behaves with them in ways they're meant to be behaved with by people.

That was a sentence worthy of a tinier brain than mine.

What I'm trying to say, poorly, with my human language "abilities," is that grabbing two things, like a soap pumper and a ball, and behaving routinely with one of them (like pumping the soap, which is what one does with a soap pumper) doesn't necessarily require as much human-like comprehension of what Sue's saying as a we might assume--that is, without thinking it through more deeply, or perhaps without trying to think like a dog or like a bonobo, or like a toddler.

Or like a gorilla--a gorilla who's arguably not a gorilla, given she's not socialized like one and given how her humans have admitted that they've intentionally raised her as a human and consider her one:



As they've raised Koko, they've narrated her mind as if it works like ours.This is evident right from the beginning of the interview portion of Morin's article. Patterson is explaining how Koko generalized the sign for "food", saying:

She would perch on this high spot where she could watch people come and go and she would sign “food” to them. It might mean “Give me the treat you’ve got,” or it might mean “I want my toothbrush,” or even just, “Engage with me.” She understood that signs had power. That particular sign got her food, so she wondered, “What else can I do with it?

She did? She wondered? And it was about how to use a sign in different settings to get what she wants? It's not just simply that it's something she could produce and did, in various circumstances? I think describing such processes as "wondering" would be giving humans too much conscious credit most of the time. But it's how we talk about minds. It's how we narrate one another's behaviors, and that of our pets, and that of whatever Koko is.

Often this is lumped into "anthropomorphism" and it's not all bad, but one could argue that it's to be avoided like the plague. If the goal is to understand how an animal thinks, then anthropomorphism risks being more of an obstacle than a helpful metaphor.

Whether the Koko project is scientifically rigorous, I have absolutely no idea. But Patterson's fantasy description of wild gorilla communication rings more of science fiction:

The free-living gorillas might talk about simple things like “Where are we going to get our next meal?” but here [at the research facility] there is so much more to talk about.

Death is one of those things. Here's Patterson's evidence that Koko understands things about death:

The caregiver showed Koko a skeleton and asked, “Is this alive or dead?” Koko signed, “Dead, draped.” “Draped” means “covered up.” 

Again, this is where my dream interviewer asks, What is the connection between "dead" and "draped"? It's even harder to understand the connection between a skeleton and draped. Can you help us understand the logic of her language? 

But we're not given that in this piece.  Morin does follow with "How would Koko know about death?" but it's not demanding a response that gets at the crucial connection between her mind and her signs.

We need to know whether what Patterson says about Koko's mind is true or not and no one seems to be able to  help us learn this. As the piece continues, my desperation for such a person escalates.

Patterson describes how Koko was making a sign that her brother made just before jumping off a rock. It took Koko's people a while to understand what she was trying to say because they hadn't seen her brother do this, but once they saw a film, it was apparent that the sign...

...means “take off” in the sense of “jump off.” Koko wanted us to take off our lab coats.

How does Koko's mind connect "jump off" a rock with "take off" your lab coats? If Morin asked this, he neither published it nor the answer. If Morin asked this question, then he is hoarding the gold all to himself.

The rest of Morin's piece is fascinating, and in parts it's heart-warming, especially if you have a soft spot for gorillas and for people who have those soft spots too. But there's still no conversation with a gorilla, or a conversation about a conversation either.

Read far enough into the article and you'll see what happened to Morin as he prepared to meet Koko:

Patterson cautioned me earlier to refrain from asking Koko questions. I was to let the gorilla take the lead. “She has that royal air about her,” the researcher explained, “and she doesn't entertain questions. Just like you wouldn’t question the queen—Koko is the same way. She’ll disengage.”

So, no conversation is going to be had with an ape, conveniently, because that totally capable ape wouldn't like it if he tried. Hm.

But there would be some lovely and touching moments through the enclosure's fence.

After recounting those, Morin reflects on what I've been discussing:

There was no way to know how much of her behavior was intentional and how much was my own or Patterson’s projection. Allegations of selective interpretation have accompanied ape-language research from the beginning. Still, it was impossible to be there interacting with her, and not feel that I was in the presence of another self-conscious being.

I don't see anything wrong with describing her as self-conscious while also doubting that her mind works like ours.  I live with a one-year old. And even before I had this baby, before I lived with an alien mind, I didn't see anything wrong with this thinking. Morin's final thoughts compare Koko's mind to an alien's, but his piece paints that alien as just another Hollywood humanoid.

When you share the link to Morin's The Atlantic piece on Facebook, the headline in the feed reads: "What gorillas can teach us about being human"

Well, what can they?

I think it's obvious from the article, and from others like it, that Koko is teaching us about our limitations. She's teaching us that we wouldn't care as much about gorillas without her fairytales; that we wouldn't care as much about these truly magnificent animals if they weren't furry humans. And that would be less depressing if we were any good at caring about actual humans.

What makes our language abilities unique? Or are they?

Anthropologists have long assumed that the more we understand of non-human primates, our closest relatives, the more we'll understand ourselves.  Anthropologists have spent untold hours observing chimps, gorillas, lemurs, baboons, and other primates in the expectation that what they'll learn will allow them to decouple nature from nurture, genetic from cultural influences on how we behave, as well as elucidate what it is about us that makes us unique: our upright posture, the size of our brains, our opposable thumbs, our language ability?

Traditional reasoning has it that tool use and abstract symbolism made our social and material world different from and superior to in a competitive sense, other species in Africa. Symboling came to involve verbal communication--language--in the way we do, that has largely been assumed to be unique and to have evolved as unique out of rudiments present, at most, in close ape relatives.

But it seems that more distant relatives can be informative as well.  In fact birds have more in common with our language abilities than do our nearer relatives in that we all, unlike our primate relatives, have auditory-vocal learning abilities.  That is, we can hear something and repeat it.  A paper in the July 2011 Nature Neuroscience, nicely summarized in Cosmos here (which, for unremembered reasons, just came up in our Twitter feed which, by definition, makes it current, right?), looks at the language of songbirds to address the question of what is unique about human language.

A paper in Science in 2002 by evolutionary biologists Marc Hauser, Tecumseh Fitch and linguist Noam Chomsky suggested that it's recursiveness, our ability to embed more and more modular bits into sentences ad infinite, that's what makes our language special.  As the Cosmos piece explained,

Recursion enables language to become an infinite system. Because clauses (e.g. "Holmes studied the footprint") can always be embedded in the next clause (e.g. "Watson said that Holmes studied the footprint") which can then be returned for the next combination (e.g. "I read that Watson said that Holmes studied the footprint"), the set of sentences and clauses that can be generated is technically infinite.

The three scientists claimed that complex, recursive syntax cannot be learned by humans or other animals, but that we must have a unique, innate specialisation for recursion and, by extension, complex syntax.

Cotton-top tamarin; Wikipedia

Fitch and Hauser followed this up with a 2004 study of cotton-top tamarins showing that they weren't able to learn recursive patterns after listening to human speech.   Then, in 2006 Timothy Gentner et al. suggested that starlings could understand recursive song, but only within pre-set structures (nicely described here, with the added plus of audio of starling songs).  This paper was challenged, however, the data said to have been wrongly interpreted.  So, back to the starting line.

And then along came Abe and Watanabe to say that songbirds really can do recursion.

We analyzed their spontaneous discrimination of auditory stimuli and found that the Bengalese finch (Lonchura striata var. domestica) can use the syntactical information processing of syllables to discriminate songs. These finches were also able to acquire artificial grammatical rules from synthesized syllable strings and to discriminate novel auditory information according to them. We found that a specific brain region was involved in such discrimination and that this ability was acquired postnatally through the encounter with various conspecific songs. Our results indicate that passerine songbirds spontaneously acquire the ability to process hierarchical structures, an ability that was previously supposed to be specific to humans.

A similar challenge has been made to this study, however, as to the 2006 Gentner et al. study, namely that there are alternative explanations for how these birds discriminate sound.  And anyway, what happens in the songbird brain is different from what happens in the human brain when these sounds are processed.

European starling; Wikipedia

But as with many many biological traits, there can be many pathways to the same trait, whether it is a morphological or a behavioral one, so the idea that only when the neural/functional pathway is the same can we say that songbirds and humans have the same language abilities is patently a non-starter.  If songbirds can do it, they can do it however they get there.

There is a lot going on in this debate, and though it's not our area, just from reading the papers and reports about the papers it seems that people are taking sides, digging in their heels and having a hard time agreeing on what constitutes evidence and falsification.

But, as with all human traits, our language ability did not evolve from whole cloth.  It has its origins in earlier traits, many of which we share with other lineages.  And it's complex and involves our complex ability to speak, our brain's complex ability to let us speak and to process sound and abstract ideas, as well as our open-ended ability to learn.  And there's that window of opportunity in young children that allows language to develop more easily than when we're older (which is why Gentner chose next to study young starlings).  We may not have much in common with birds in terms of how we produce sound or the structures of our brains that process it, and indeed we're much closer to non-human primates in that, but this may be insignificant.

If it's true that songbirds do have a sort of recursive language, as we do, it may well have evolved separately from our own ability; an example of a kind of convergent evolution.  But what this whole story means to us is that the idea that a single trait can be plucked from the highly complex mix of traits that is language, which involves so many different parts of our anatomy and our brain, and can be used to define what it is that makes our language abilities unique is, to us, not a very useful one.  And, if it's possible to contest the uniqueness of any trait that is said to be the one that makes our language special, this is a good indication that it's not a single trait at all, but a whole suite of things that, added up, give us the ability to argue about this at all.

This in no way minimizes the interest or importance of understanding how symboling and language have characterized or shaped our species during its evolution, nor of the complex neural mechanisms that must be involved.  But it does show that it may be the trait, and not some particular mechanism, that is what is important, and that we should examine language on its own terms--in whatever species--rather than be too careless in human exceptionalism.

Hurling words and turds, an evolutionary link

Humans are excellent throwers. Even poor throwers, or famously ridiculed ones, are still pretty skilled compared to other species with grasping hands. We're so good at throwing things, it's hard not to wonder why. 

This is when you say, "Our arm anatomy, dummy. That's why humans are good throwers."

Our shoulder, arm, wrist and hand anatomy is mostly very similar to other primates' but the differences are critical to our ability to throw so well. (See here for more in-depth discussion of head-to-glutes-to-toe throwing anatomy.)  Once the arms were freed from their locomotor role their anatomy could respond to different selective pressures (or not) and one of those pressures might have still been locomotion (arm swing) and another was likely throwing, considering the benefits of action-at-a-distance for obtaining food, avoiding predators, and interacting (not so nicely) with other humans. 

But it's not all about the arm. Insert a chimpanzee's brain into a human's body and it'd be a safe bet that Frankenzee couldn't throw like Frank Reich... or any of us. That's because throwing well by human standards isn't just about human limb anatomy, it's about controlling that human limb anatomy with the human brain. 

Some popular and well-supported explanations for the origin of throwing are brain-based. The coordination of the body's movements and timing of those movements relative to the distance and velocity of a target is nothing less than genius. This same sort of coordination is required for language which is why hypotheses compare and even link throwing evolution to language evolution. Bill Calvin fleshed out an idea in The Throwing Madonna where he hypothesized that throwing enabled the evolution of language because both are controlled in large part by lateralized functions in overlapping regions (Broca's area) of the left hemisphere.

So it's only because of our big and specialized brains that we can both talk trash and sink a clutch three-pointer.

There's a field of research on understanding the biology, biomechanics, and physics of throwing behavior but only a subset of that research is rooted in an evolutionary, comparative approach. (For one good example, see Neil Roach's work.) As mentioned above, this is because few other species actually throw. However, lucky for us and despite their lackluster ability, chimpanzees do love to throw s--t.

This week some researchers who are interested in the neurobiology of throwing published a paper that made use of this hilarious habit--made famous the world over by zoo visitors with cameras and youtube accounts.

Like the study's authors, if you hang out around chimps long enough, their variation just screams out at you. For one, they vary in how they scream out at you.  But they also vary in their penchant for throwing. And this observation might cause you to ask yourself, If throwing is a brainy activity are the brains of the chimps who like to throw s--t any different from the brains of chimps who'd rather not? And if they're different, how are they different? What parts of the brains are different? Is there anything else about the chimps who like to throw s--t that separates them from those that don't?

These are the questions that Hopkins and his mates asked in their paper.

First they divvied up 78 chimps at the Yerkes National Primate Research Center into those who "reliably throw" (poop and food, mostly) and those who do not. I could not find more details on how this distinction was made, but if you spend your days at the YNPRC it's probably pretty obvious who the' THROWING+' and the 'THROWING-' are. (And you'd probably best learn fast before the s--t hits the man.)

Then they anesthetized each of them so they could MRI their brains to test for any brain anatomy differences.  (This was hopefully, and probably legally bound to be, coordinated with a necessary physical exam for each of these animals so the trauma was for good, health-related cause.)

To test for behavioral or personality differences, they performed the 'Primate Cognition Test Battery' or PCTB (Herrmann et al., 2007) on each of them.

When they compared the brain scans, they found that THROWING+ chimps had significantly more white matter relative to grey matter in the inferior frontal gyrus, which is the homolog to Broca's area in humans. And they also had significantly more white matter in the motor-hand area of the precentral gyrus, which is associated with handedness. Increased white matter is important because it indicates more myelinated interneurons that connect different cortical regions, suggesting to the authors that "learning to throw may alter the connectivity between premotor and primary motor cortex in the chimpanzee." (p. 44)

For an adaptive hypothesis in humans to come from this we'd need to parse out causes and effects. (The authors acknowledge this issue that haunts so many comparative studies.)

When they compared the results of the PCTB the only significant difference was that THROWING+ chimps scored higher in "communicate" points.

Hopkins et al., 2011

As far as I can tell, this "communicate" score comes from tests of a task called "comprehension" where experimenters gaze and point at targets and apes are assessed on how they respond, and another task called "production" where apes are assessed on whether they produce communicative signals (such as manual gestures) to indicate where food is hidden in hopes that an ignorant human will find it and give it to them.

This all seems so far removed from chimpanzees throwing s--t doesn't it?

Especially considering that the authors discuss these findings in support of the connection between language and throwing in human evolution.

Despite all the dots that need to be connected, throwers' brains did have more white matter in potentially significant (to throwing and language) centers of the brain and throwers did outperform non-throwers in "communicate" tasks on the PCTB. How else do you explain these things without throwing as part of the explanation, and maybe significantly so?  And if you're on board with that, how about throwing as a critical precursor to language... as the means for laying down the neural tracks that were later used for language?

Evolution of one wouldn't have occurred without evolution of the other one first. Why not? This isn't so scandalous. Just about everything else exists because of what came before. 

The non-humanness of human language

The aspects of human language that make it uniquely human remain an open question, but the short answer is that there isn't as much that's unique as had once been thought. That is, it seems that much of the foundation for our language perception and speech abilities was laid down long before the evolution of hominids, and is shared by even distantly-related species. And, a surprising lot about language that seems uniquely human seems to be learned rather than innate.

We've blogged here before about birds and cows, among many other non-human purveyors of sound and meaning, having regional dialects, about non-human primates, and dogs, learning and responding to basic human language, and human newborns seeming to cry in a way that reflects their native language. Language acquisition and its evolution are very active areas of research. The results of two studies published last week are further evidence of language abilities that aren't ours alone.

A paper published in the Proceedings of the Royal Society B (Zebra finches exhibit speaker-independent phonetic perception of human speech, Ohms et al., paper published online Dec 2, 2009) reports that finches are able to discriminate between two vowels in single-syllable words, independent of the speaker.

There is an ongoing debate about whether the ability to form phonetic categories that underlie such distinctions indicates the presence of uniquely evolved, speech-linked perceptual abilities, or is based on more general ones shared with other species. We demonstrate that zebra finches (Taeniopygia guttata) can discriminate and categorize monosyllabic words that differ in their vowel and transfer this categorization to the same words spoken by novel speakers independent of the sex of the voices. Our analysis indicates that the birds, like humans, use intrinsic and extrinsic speaker normalization to make the categorization. This finding shows that there is no need to invoke special mechanisms, evolved together with language, to explain this feature of speech perception.

And, a report of a long-term study of vocalization among Campbell's monkeys in the Ivory Coast, suggests that these animals are able to combine sounds to make new meaning in response to external events, such as the sighting of predators (Campbell’s monkeys concatenate vocalizations into context-specific call sequences, Ouattaraa et al., PNAS, published online Dec 9, 2009). "These call combinations were not random, but the product of a number of principles, which governed how semantic content was obtained."

The trend seems to be toward a chipping away at specifics that have previously been thought to make human language a singular attribute of our species, or at least as singular as many would fancy it to be. It seems likely that there's no one aspect of our language and perception capabilities that can explain how we alone have the ability to give abstract meaning to sound or to convey completely new ideas between ourselves in the open-ended ways that we do.

Language is another complex trait that can't be explained by reducing it to its many parts -- the use of prefixes, suffixes, the effect of a single gene, our sound discrimination abilities, and so on. Instead, it's an emergent property that flows from what our brain allows us to make of the world, combined with our biological ability to make and detect sound (although, that's clearly secondary and not essential, as the complexity of sign language used around the world demonstrates), and built on a foundation that has been evolving for millions of years. Hundreds or thousands of genes are required for this, as the plethora of mutations in genes that affect cognitive abilities including language clearly show.

This makes sense. Every trait evolves from precursors. Every step of the way we humans are shown not to be unique but to be more a part of the Nature that produced us -- even if, albeit, every species is, almost by definition, unique. It is hubris to think otherwise, be it with respect to language or even consciousness. That doesn't take away in any sense from the interesting question of what humans are and how we got that way -- and why we are as much different from other species as we are.

And, of course, if anything, it is more evidence against those who cling hopefully to the idea that we were specially created outside of Nature rather than evolved within it. Each discovery of the role of environments in molding our basic abilities, or of characteristics we share with other species, even very distant ones, confirms our place in Nature. And it stimulates further interesting research. But it won't lead to easy categorical conclusions.