Survey of Terrence Deacon's "The Symbolic Species",
by jjackson@interalpha.co.uk



Chapter 12: Symbolic Origins


Having considered the path of symbolic progress, we now consider why it all started in the first place, seeing, as Deacon claims here and elsewhere in the book, that it's necessary to learn a lot of indexical associations that don't pay off until they are all learned. [I'm not so sure about this. I suspect a detailed working model of symbolic structures could be developed that paid off at all stages during the constructiuon of a symbol.]

He couches the issue in terms of a new and at that time still rather dodgy alternative form (symbolic) of communication, enjoying special utility in a particular social context. As sexual selection tends to produce significant changes in communication in other species, he posits it as the anvil on which the proto-human language was first forged. [I am very wary of explanations based on sexual selection; it seems there is no accounting for taste in that area, so it makes a magical last resort, often used to explain odd features for example in dinosaurs: "Birds grew feathers for sexual selection; Triceratops grew horns for sexual competition" etc etc etc. I strongly suspect that features important in sexual selection arise through simple extensions of signs of "fitness" in a mate, and once the sign of sexual excellence departs to far from the nitty gritty, it loses its importance. However, the symbolic ability isn't used in that facile way as a sign in Deacon's explanation...]

Why a special social operator such as early language would have been particularly valuable for the advanced Australopithecines was as follows:

For reasons many of us would be prepared to accept, and which he describes at length in the book, the proto-habilines were able to benefit from meat consumption through hunting (or perhaps scavenging) if they foraged in male groups, where females, particularly mothers, could not usually participate.

(In passing, he suggests that the reason our ancestors had smaller canines than normal in primates (and even chimps) was because they ground tougher plant matter than primates with large canines, which would get in the way of all the milling. Therefore, to transfer to meat they needed tools (weapons).)

Pack "huinting" for meat interacts in an interesting way with the unusual human formula of fairly stable male-female pair bonds, and pack hunting. It is unusal because the males have the problem of possibibly being cuckolded while they are out, and the mothers have the problem of the fathers possibly meeting another female and sharing the fruits of the hunt with her instead while out galavanting (indeed in more recent historical times, this seems to have been part of the appeal of hunting!).

The various social hunting arrangements of other pack-hunting carnivorous mammals seem designed to get round these problems, and when they do, specialised social communication plays a key role - often chemical communication. As primates, we are not well suited to this mode.

In chimps the incentive to give meat to nursing females is lessened due to uncertain paternity. For hunting to benefit females and young (and if it became an essential part of early hominids' economy it would have to) a guaranteed, unambiguous and egalitarian mating mode would have to be contrived. To cut a long story short, this particular instance of pair-bonding, particularly when accepted as such throughout the group, and not depending on past experiences between members of the pair and other individuals, is presented as a symbolic entity. "Marriage, in all its incredible variety, is the regulation of reproductive relationships by symbolic means...". It is claimed that indexical mechanisms cannot do this.

[Part of the reason I can't concur is a suspicion that Terrence is perhaps taking every high level concept as a symbol. I believe animals show evidence of implementing entities such as a class of behaviour or an emotional state, which arise as a particular compound of certain precursors, yet which do not represent anything external. In other animals these are instinctive (as they can be in humans too) and as such are not created afresh during an individual's lifetime but are evolved in the old fashioned way. I see no reason why marriage and certain social procedures arising from it couldn't have arisen in this way (and I would think that when we teach geese to talk they will agree :-) ).]

The protohumans would have needed considerable support in order to establish their symbolic constructs, just as Kanzi et al did. Ritual is advocated to fill this requirement, and Terrence draws our attention to the frequency of this curious feature throughout human culture. "Thus, many ritual activities take the form of an ideal symbol discovery process."

[If marriage (and other social concepts) need to be known absolutely and not learned by the rest of the group for each couple, what about all this "impressing the symbol through ritual" ? Presumably the learning process is not "done away with" after all, merely got out of the way all at once before people have a chance to learn through their mistakes.]

"In general, it is probably fair to conclude that the more difficult the social symbolic problem - either because of the intrinsic conceptual difficulties, or because of the intensity of the countervailing social forces that must be mediated - the more likely that highly ritualised means will be required to establish appropriate symbolic responses." p403. [I love this quote, partly because of all the Highly Respectable carry-ons this covers that one really couldn't mention lest one be considered Peripheral to Normal Society, partly because of all the desirable concepts (social and intellectual) that won't become mainstream anytime soon because of "countervailing social forces", but not least because Terrence's thesis here will itself find "intrinsic conceptual difficulties" and those countervailing forces serious obstacles to its own acceptance.]

One social symbolic relationship he mentions is peace, and he describes the way it is established as a symbol through a collection of consituent indexes, Yanomamo style. Promises just won't do on their own; genuine physical gifts and feasts representing - indeed typical of and defining - peace, are exchanged. [One might say "a promise is only a promise, but a good sandwich is a meal - even if given to you by a liar and a hypocrite".] Also, situations are presented where choices between peaceful and warlike behaviour must be made, which will have more power than a simple statement. It has been said that to refuse an invitation to the feast is a declaration of war.

Such rituals are what Terrence has in mind when he says that vocalisations would have played a minor role in the earliest symbolic communications. "Probably not until Homo erectus were the equivalent of words available."

His diagram on page 409 shows various significant developments starting (mostly during Homo habilis) at (dates approx):

2.4 mya: Stone tool making and group hunting
2.16 mya: Brain-restructuring for speech & symbols
2.16 mya: Male provisioning, pair bonding & mating contracts
1.7 mya: Laryngeal descent & syntactic complexity

(His borderlines at least in our lineage are approximately:
Australopithecus -> 2.4 mya -> H.habili -> 1.8 mya -> "Homo erectus".)

He suggests the original vegetarianism and polygyny of Australopithecus dwindled during H.h as the other developments mentioned above progressed.

[Two comments that might be made on these timetables: if group hunting preceded mating contracts by something like a quarter of a million years, there must have been an awful lot of tension amongst them during that gap. Also, if the usefulness of mating contracts was the spark to symbolic development, I find it curious that tool use preceded this by 0.24 mys and had no role in symbolic processing.]

[Instead, I think tools would have provided the incentive for symbolisation. How often has your day been ruined because you've left something vital at home? "Things" tend to make your life more complicated through considerations of use as well as manufacture. Say for example the band had to travel to a place where stone tools would be needed. They may need to obtain the flints from some other place. In order to carry them they may need to obtain stuff to make a bag from. This bag itself may give rise to further complications such as needing tools for its manufacture, or being made out of different materials at different times of the year. In order to survive in or spread across areas where ordinary animals could not, it would somethimes have been vital to manage these extra complications adequately. The intellectuality implicit in stone tool making is only a small part of the picture. For every stone item a population used, there will have been - how many - four? forty? different perishable tools in use. The sheer complication of life with a tool kit would require considerably more planning and communication. One can only sympathise at the task of one of a group of Australopithecines on their way to becoming habilines having to remind another to stop off at the creek for some pebbles before going on get the food.]



Chapter 13: A Serendipitous Mind


Here we consider the way in which even humans disadvantaged by eg injury retain a considerable fraction of their lingual abilities, yet even the most intelligent animals strain to make any approach towards human levels of symbolic ability. Terrence points out the "over engineering" that must have evolved to ensure as little chance as possible of any human coming into the world without their essential human characteristic. "Language acquisition had to become failsafe."

[However, in so doing, we are posed with an interesting issue: if symbolic processing is merely the fuller utilisation of higher level concepts (concepts compounded from lower level ones, further distanced from raw external or internal perceptions/actions, and able to overpower rules based on lower level slots), then what is it about "intelligent animals" that makes them more so than others? An emphasis on higher level concepts but short of the human degree? An ability to induce rules from a wider variety of basic components though still at a low level? To be fair, it should be said my definition of symbolic processing might well be disputed by Deacon, though he does describe the learning of higher order conditional associations as the central learning problem[..] that makes symbolic associations so hard for other species.]

However the special cognitive emphasis this results in has effects in areas other than language, particularly since it is such a basic non-specialised speciality, and not solely a language enabling mechanism. Deacon warns us throughout the book that it is dangerously unreliable for us to describe either other extant species' or our ancestors' minds in terms of simple variations on ours.

He pays much attention to aspects of prosody (tone of voice etc), which he suspects may be particularly associated with the less-verbal side of the brain (usually the right), and also the non-verbal forms of utterance controlled subcortically in other animals but to a lesser extent in humans. He points out that we actually have fewer characteristic non-verbal "calls", understandably since language does the job better, but laughing, and sobbing/crying had to be retained for their use at the pre-verbal/pre-symbolic stage. Laughter, referred to as a vehicle for social bonding and a play signal, has leaked out to cover aspects that didn't really occur pre-humans. Humour often appears to be due to the discovery of alternative conceptual structures, a process very much to do with symbolic operations, and it is interesting that the successful completion of a hierarchical symbolic structure [perhaps a representation of reality], should be associated with pleasure. [Perhaps it also represents an implicit earlier error in the interpretation of structure, but a non-serious one, so that the "play signalling" aspect of laughter might be involved.]

Deacon gives other examples of successful symbolic operations giving pleasure, typically problem-solving. He talks about reinforecement applying under these circumstances [gratifying for those hypothesising the assembly of concept structures being strengthened by "successs"]; amongst other things he also speculates on the "release" of laughter reflecting the disengagement of pre-frontal control, and a public symptom of a resolution of internal conflict.

He ponders the question of what it means to read another's mind, and whether we may have an advantage over other animals in this. He suggests three ways of bringing multiple minds into concert: mimicry (iconic); reaction to a common external stimulus (indexical); modelling the other's mind (symbolic). However he does suggest animals such as pets are very good at judging our moods and certain intentions.

In the sort of area the book is moving into now, propositions become very hard to justify, particularly since we still really don't have a very detailed definition of "symbolic", and this is not the kind of area where I would want to make statements about what cannot happen. I am not convinced by Terrence's conviction that nothing symbolic can be instinctive, in this instance that ethical considerations, which he says are a result of symbolic processes, cannot be innate. He points out that we can be emotionally aroused by symbolic operations (eg stories) but adds that we are therefore the only creature that can experience a genuine conflict of simultaneous emotional states. Would he deny this predicament in an animal in which hunger and fear were battling to decide the course of action to be taken? Perhaps he means two complete opposites together, such as hungry and not hungry; afraid, and, because of the story being listened to, not being afraid.



Next section of survey