Let me preface this by saying that Incomplete Nature is probably one of the most daring and original published scientific monographs I’ve ever read. Of course, it could also be one of the worst, it's actually impossible to tell. That said, I’ve had Terrence Deacon’s first book, The Symbolic Species, sitting on a shelf at my house for about five years now. I picked it up when I was a sophomore in college, at the second link of a five-year chain that went evolutionary biology → evolutionary psychology → cognitive neuroscience → philosophy of mind → consciousness → causality and information theory → oh-my-God-nothing-is-real. At the time I clearly wasn’t ready for the book, as I read about the first ten pages before putting it down.
This was for two reasons. The first is that his books are long, dense, and convoluted, and his sentences tend to loop-the-loop back on themselves and self-cannibalize. This is certainly true for Incomplete Nature, which clocks in at over 600 pages, and has entire sections that are interesting and informative but completely unnecessary for the main point/plot (see also Melville 1851).
The second reason is that Terrence is a little bit, well, verbose. And by verbose I mean logorrheic, prolix, pedantic, etc etc. He also makes up his own words - a lot. Like Shakespeare, or Sarah Palin. This habit is clearly on full-throttle in Incomplete Nature, but mostly that’s because what he’s trying to talk about is something that is so mind-bendingly different from the corpuscular metaphysics of Descartes that most scientists and lay-people take for granted. So for example, in Incomplete Nature, he makes up the word “morphodynamics” before discovering it’s been in use for about eighty years in embryology and geology (cue five page digression into the history of the nominally similar term that’s not quite the same as the term he had in mind but is still kind of the same, or same enough to devote five pages to, followed by how his definition is different. Just give up the word, Terrence - let them take it). Other made-up words include “absential”, “autogen”, "autogenesis”, “constitutive absence”, “contragrade”, “ententional”, “entropy ratchet” and I’m nowhere near the end of the glossary so I’ll stop. So why all the neologisms? Just what point is Terrence trying to get across that requires forging new words just to get our attention?
Terrence Deacon sees a problem in our sciences, and that problem is this: how can information, meaning, representation, function, consciousness, have causal powers? As human beings we naively use these terms all the time, and our causal explanations of the world work so incredibly well! Let me put it in an example (as Terrence would approve of). So I know that tomorrow my morning latte will have a little snowman made of milk on the surface because my barista and I have been flirting for weeks and I’m pretty sure she likes me. That high-level pattern of flirtatious behavior is incredibly predictive of the future complicated design of the snowman, but the predominant (if ill-defined) position is that causality is all “happening” down at the atomic level. Therefore such higher-level patterns are meaningless - good for compression in stupid human brains, but not *real*. Terrence disagrees, and he disagrees in a number of ways. He wants to say that not only did consciousness cause the snowman to assume complex shapes, but that further past events cause the snowman, events that no longer even exist (and thus are “absential”). He wants to bring back teleology, show how life can beat entropy, show how evolution creates function, show how... well, okay, it’s a lot. If these problems seem so wide-ranging as to be almost all encompassing, it's because they are. Terrence goes from the problem of the origin of life (how do anti-entropic traveling waves of complexity get constructed?) to the origin of consciousness (what is the simplest form of a self?). Due to the size of the book itself, I will only be a describing it as abstractly as possible - a difficult task, as each portion builds on the last, incorporating the necessitated vocabulary. Or, to paraphrase David Foster Wallace in an interview I once saw: “If I could sum up the theme of the book I wouldn’t have had to write it.”
So I’ll just tell you the conclusion of the book, and then I can get on to more interesting things, such as why an anthropologist is writing a book about how to beat entropy. The conclusion: consciousness and the self are teleodynamic processes defined by the informational constraints they place on the dynamical phase-space of the organism by their anti-entropic (and thus function-creating) nature. Being anti-entropic means being fundamentally incomplete (kept from resting level). That's fine, as long as you've defined all your terms, equations, shown how it operates in simple example systems, etc, but Terrence seems to operate in the style of "intellectual jazz" - reading Deacon is like reading Sartre: "the being of the being of the becoming... of the being." The problem is that this all flirts so close to complete semantic vacuity. Mathematical descriptions are sparse, just a handful of them, and strict definitions are nowhere to be found. For example, the Belsousov-Zhabotinsky reaction is a chemical reaction in a shallow dish that demonstrates self-organization far from resting state. Is it conscious? Teleological? What level of the "entropy ratchet" is it on? I doubt Terrence has any answers. I can make out his primary interest, which is that the notion of information (which is connected to entropy), which Terrence believes needs to change in order to incorporate consciousness, and with it meaning, semantics, and function. That's what I'll address from now on.
Information theory as a key to consciousness has a pedigree - David Chalmers (1996) admits thinking that such a theory could be constructed, and Giulio Tononi (2010) actually did construct one: an info-theoretic measure of consciousness that is fully formalized in discrete systems. But why is Terrence Deacon, an anthropologist, and Tononi, a psychiatrist, the only ones following this up? One dismissive reason would be that they’re crackpots, who don’t really understand either information theory or physics. The other, more interesting reason, lies in how scientists have performed reduction. This is most well put by John Searle (1992), where he points out that, throughout history, reduction has involved confinement of various properties of the phenomenon under study to the subjective. For example, the redness of red is said to be ‘in the perspective’ and not out in the world - this allows us to comfortably come up with a scientific definition of red (photon emissions at 600 nanometers). I think that this kind of reduction (a push back into the head) occurred in Claude Shannon’s work on information theory, and that, when we use information theory to look at the brain, we notice a troubling paradox.
Information, in the Shannon sense, is not really an ‘intrinsic’ quality - rather instead it is observer-defined, based off of a vocabulary agreed upon before hand. Shannon's Entropy function is a measure of how surprised one observer is, given that another observer sent a signal. But then we look in the brain and see... the same kind of observer-defined information. Shannon’s separation of information from meaning and consciousness has lead to an elimination of meaning from science. This is fine for those whose jobs involve building better telephone wires and computers, they don’t care that what they’re doing is based off of prior agreement on vocabulary between observers. Even mathematicians and physicists aren’t bothered. But it is neuroscientists (in broad definition), the people who actually have to hold a brain in their hand, or listen to the pop-pop of in vivo neurons, that are really struck by the paradox. Terrence’s book is, at its heart, an attempt to solve this paradox, and bring the ineffable back in the game. Of course he fails in the sense that the majority of the ideas in the book are descriptive and not formalized, at worst incoherent, at best incomplete.