From Philosophy to Neuroscience and Back Again. An Interview with Terrence Deacon. Part II

From the Symbolic Species to Incomplete Nature

Auguste Nahas 

So it’s only in your first book, The Symbolic Species, that you came back to Peirce. 

Terrence Deacon

Yes. I was thinking about the evolutionary origins of human language. It was very clear to me that you just couldn’t talk about ‘pre-language’. Animals aren’t speaking in a simpler form of [human] language. There’s no such thing as bird language or cat language in that sense. It’s a really different kind of communication. And yet, there must be some kind of continuity between human language and what came prior to it. The semiotic perspective helps you see the continuity and discontinuity. If our human ancestors were communicating without language, then we need to think about language as a special case of a much larger phenomenon, which is (bio)-semiosis: the capacity to interpret and generate signs. The discontinuity can then be framed in terms of the Peircean idea that there are different types of signs, and that each type requires distinct capacities to generate and interpret them. The Symbolic Species is about how our brains evolved to interpret and generate a highly unusual type of sign: the symbol. I realized that we need to rethink evolutionary theory to talk about how this unusual form of communication evolved. So the symbolic species really sort of brings together for the first time these different threads. But it took a decade for me to flesh this out.

Auguste Nahas 

Speaking of continuity and discontinuity. What strikes me as a real connecting point between The Symbolic Species and your second book, Incomplete Nature, is this idea that if there’s no “pre-language” in non-human animals, but there’s still something like communication happening. So getting into view precisely what is similar and what’s different is the key. And I get the sense that this is really the theme behind Incomplete Nature as well, but applied to life as a whole. What distinguishes living things (biology) from the non-living world.

Terrence Deacon

Yes, the Symbolic Species set me up to think about the problem of emergence. And that’s because one of the things I learned from my neuroscience research in the late 80s, is that it is very clear that the human brain has no new parts that aren’t found in monkey or ape brains. Brain evolution is incredibly conservative. Things enlarge and shrink, but on the whole, structures don’t change much. So it seemed possible to predict human brain circuits critical for language from monkey brains. Now this raises an important question: how is this possible, if these are species that didn’t have any capacity to do anything like language? This is the continuity and discontinuity question from the neuroscience perspective: how could it be that we are continuous with these other species in terms of brain anatomy, and yet that we are doing something radically and fundamentally different semiotically? People still don’t understand, I think, the semiotic uniqueness of language. You just don’t find in linguistics people talking about anything semiotically, in this sense. They talk about language, and then behavior. So they’re treated as separate things, it’s almost a sort of Cartesian split. And what I realized is that though Peirce has carefully described the distinction between symbols and other kinds of referential relationships, the term symbol had since then become compromised, especially over the course of the previous century. So that symbol really now simply referred to arbitrary marks.

Auguste Nahas 

Do you mean a computational understanding of symbol?

Terrence Deacon

Yes, although it has an origin in the history of symbolic logic. Basically, symbolic logic gives rise to computer theory. And as a result, we get this idea that symbols are signs that refer arbitrarily. This allows you to talk about it like it’s a code. You got a bunch of marks, and you map it to a bunch of other things. That’s an arbitrary code relationship, and that’s why we use codes for encryption –  there’s no natural way to make this mapping. The idea at the time, and it’s still, I think, in the background of most thinking about language is that this kind of symbol is the most basic type of sign. From this perspective, iconic signs – which refer through likeness – and indexical signs – which refer through spatiotemporal contiguity – are more complex, because they require something added. This is, in effect, a complete – and in my view, mistaken – inversion of Peirce’s sign typology. It led to people like Stevan Harnad coining the ‘symbol grounding problem’ in the 90s: how does a symbol come to map to such and such an arbitrary meaning? Through what process does it become ‘grounded’? But notice how this question is totally inverted. From a Peircean perspective, if symbols are constructed from simpler (iconic and indexical) signs that refer non-arbitrarily – i.e., are inherently grounded – then the question is about how you use those signs to construct a new type of sign (symbols) which allow reference to become ungrounded. So this inverted way of thinking about symbols was – and still is – a real problem. I tried to address it in the Symbolic Species, where I argued that symbols are the exception, not the rule. They are constructed from simpler iconic and indexical signs. Unfortunately, I think the work fell on deaf ears in the linguistics community, many of whom thought that stimulus-response behavior is a kind of symbolic reference in virtue of its being arbitrary. That being said, it was well received by neuroscientists and evolutionary biologists. 

Auguste Nahas  

But it took a while for the symbolic species to gain recognition right? Didn’t it end up winning an award many years later?

Terrence Deacon  

Yeah. Almost a decade later. It was judged to be the most influential book in anthropology. So I gained a major prize there. And it still sells pretty well, 23 years later.

Auguste Nahas

So it took over 10 years to go from the Symbolic Species to your second book, Incomplete Nature, which is ostensibly more philosophical and talks about emergence of life and mind. That may seem like a major shift to readers. What happened during that time?

Terrence Deacon

During the time that I was writing the Symbolic Species, I was working at Harvard Medical School with researchers who were doing transplantation research of brain cells across species, in order to better understand Parkinson’s disease. I got involved, in part, in the hope that it would help me understand similarities and differences in species function. My colleagues and I developed distinctive molecular markers, which allowed us to see a single pig cell grow, develop and make connections in a damaged rat’s brain. The cells found appropriate targets in this totally other kind of brain! And because pig brains take longer to mature, the pig neurons had an advantage. They were able to, in a sense, spend more time exploring the rat’s brain and finding appropriate targets. So they did a better job of restoring normal function in the rat brain than rat cells transplanted from healthy brains. We learned a lot from that, particularly the evo-devo perspective on brains. That is, to understand how brains evolved, you need to understand how they develop. The Symbolic Species, especially the middle part of that book, takes a lot of insights from that work. So it brought many strands of my thinking together.

Andrea Gambarotto

Could you say a bit more about how the neurological story meets the semiotic story?

Terrence Deacon

Well I realized that though very different parts of the human brain became recruited for language processing, they all had precursors in monkey and ape species, even if their connections and size were different. That’s what sort of is at the key to the symbolic species argument: what’s so different about symbolic thinking? What changed in the human brain? What makes human brains stand out differently from other brains? How does the change in the brain correspond to the change in processing demands that symbols provide? So this is a question about the emergence of a radically new function, that as it began to emerge, began radically restructuring human brains. In simple terms, the emergence question is: how does the same old stuff begin to do something fundamentally different? That’s basically what the subtitle of the book “The co-evolution of the brain with language” is saying: the demands of symbol processing are what caused human brains to diverge in their structure and function from other brains. But this happens all the time in evolution. So to my mind, some of the most important and interesting evolutionary transitions are these major emergent transitions, and these days I’m working on a book about how there’s a basic pattern that underlies these evolutionary transitions. But the first and most important emergent phenomenon, which starts it all, is the emergence of life itself. And that’s where Incomplete Nature comes in.

Auguste Nahas

So the parallel between the two books is that in the evolution of the human brain, we’re trying to understand how despite material and biological continuity, there is also the rise of a radically new function. And in the origins of life, though there is a kind of material-energetic continuity, there’s also something radically new.

Terrence Deacon

Yes, that would be the parallel, but there’s also an important difference to bear in mind. In the symbolic species story, it’s symbolic communication that comes first. It’s not that the brain changes and then symbols become possible. But insofar as Incomplete Nature is a story about the first system that is capable of semiosis, that can’t be the case. There’s no semiosis, and therefore no signs, prior to life. In human brain evolution, we know symbols are crudely possible in species that don’t have human brains – and we’ve shown this in chimpanzees and bonobos. And to some extent, in other species like dogs, we can get some little bit of symbolic capacity but it takes a lot of effort to do it, because [symbol use] requires brains to function in a very different way. But when symbolic communication becomes critical for a species – and the last part of that book discusses how communicating by symbols provides something that no other form of communication can provide – you get a selection force restructuring the physical nature of brains to deal with this communicative problem, this new ‘symbolic’ semiotic problem. 

Auguste Nahas:

Interesting how that’s so radically anti-Cartesian. It means that what Descartes thought was immaterial – what you would call symbolic semiotics – had a profound physical effect on our brains, and on evolution itself.

Terrence Deacon:

Absolutely. Symbolic communication changed the structure of brains, affecting biological evolution, and since then our symbol use has allowed us to drastically reshape the planet itself. The same neurological structures that had evolved prior to humans for sound analysis, visual analysis, motor control, and so on, became recruited to do something very different, and they have to do it together, synergistically. They’re forced to work together in a novel way to produce and interpret this unusual form of communication. So what kind of process produces emergence, produces radical discontinuity in function, and produces a radical discontinuity in structure? I realized that this was a way to think about what you might call “the semiotic problem”: the problem of how information in reference is linked to the physical problems of evolution. And, of course, that in another way, is the problem of the origins of life.

But before going any further I should say that I started to write Incomplete Nature as a very different book, it was going to be a more direct follow up to the Symbolic Species. The original title was Homunculus: an attempt to deal with the problem that so much of our way of thinking about brain function is as if there’s little guys in the head (homunculi). We talk about how the brain’s “doing”, “understanding”, “seeing”, and so on. But when we dig into those concepts we often find unanalyzed homuncular concepts. The book was going to focus on this. I was a few chapters into it when my computer was stolen at a conference in Zurich in 2001, and much of that work was lost.

Auguste Nahas:

That’s every writer’s worst nightmare. But did that chance-happening actually influence the re-rewriting or rethinking of the topic for the book?

Terrence Deacon:

Completely. As I began rewriting, I realized that I wasn’t getting straight at the underlying issues. So I decided to focus on those first, and end the book with my reflections on homuncular thinking in neuroscience. But then that first part took over and I found myself never really getting to the brain, so it ended up being a sort of prequel to the Symbolic Species. I wanted to talk about the concepts of emergence, information, self, interpretation, and teleology – all concepts that were critical to my account of language and evolution, but largely implicit in the Symbolic Species. That began in the late 90s and early 2000s, and you get a sense of where I was going in the chapters I wrote in a book on the Baldwin effect reconsidered (Deacon 2003a; 2003b).

Auguste Nahas

So Incomplete Nature seems like a real return to the more philosophical, or metaphysical aspects of Peirce’s work.

Terrence Deacon: 

Yes, it sent me back to all of that.

Parti III. “Beyond Incomplete Nature” coming soon, stay tuned.