Visions of the Semantic Apocalypse: James Andow and Dispositional Metasemantics

by rsbakker

The big problem faced by dispositionalist accounts of meaning lies in their inability to explain the apparent normativity of meaning. Claims that the meaning of X turns on the disposition to utter ‘X’ requires some way to explain the pragmatic dimensions of meaning, the fact that ‘X’ can be both shared and misapplied. Every attempt to pin meaning to natural facts, even ones so low-grained as dispositions, runs aground on the external relationality of the natural, the fact that things in the world just do not stand in relations of rightness or wrongness relative one another. No matter how many natural parameters you pile onto your dispositions, you will still have no way of determining the correctness of any given application of X.

This problem falls into the wheelhouse of heuristic neglect. If we understand that human cognition is fractionate, then the inability of dispositions to solve for correctness pretty clearly indicates a conflict between cognitive subsystems. But if we let metacognitive neglect, our matter of fact blindness to our own cognitive constitution, dupe us into thinking we possess one big happy cognition, this conflict is bound to seem deeply mysterious, a clash of black cows in the night. And as history shows us, mysterious problems beget mysterious answers.

So for normativists, this means that only intentional cognition, those systems adapted to solve problems via articulations of ‘right or wrong’ talk, can hope to solve the theoretical nature of meaning. For dispositionalists, however, this amounts to ceding whole domains of nature hostage to perpetual philosophical disputation. The only alternative, they think, is to collect and shuffle the cards yet again, in the hope that some articulation of natural facts will somehow lay correctness bare. The history of science, after all, is a history of uncovering hidden factors—a priori intuitions be damned. Even still, it remains very hard to understand how to stack external relations into normative relations. Ignorant of the structure of intentional cognition, and the differences between it and natural (mechanical) cognition, the dispositionalist assumes that meaning is real, and that since all real things are ultimately natural, meaning must have a natural locus and function. Both approaches find themselves stalled in different vestibules of the same crash space.

For me, the only way to naturalize meaning is to understand it not as something ‘real out there’ but as a component of intentional cognition, biologically understood. The trick lies in stacking external relations into the mirage of normative relations: laying out the heuristic misapplications generating traditional philosophical crash spaces. The actual functions of linguistic communication turn on the vast differential systems implementing it. We focus on the only things we apparently see. Given the intuition of sufficiency arising out of neglect, we assume these form autonomous systems. And so tools that allow conscious cognition to blindly mediate the function of vast differential systems—histories, both personal and evolutionary—become an ontological nightmare.

In “Zebras, Intransigence & Semantic Apocalypse: Problems for Dispositional Metasemantics,” James Andow considers the dispositionalist attempt to solve for normativity via the notion of ‘complete information.’ The title alone had me hooked (for obvious reasons), but the argument Andow lays out is a wry and fascinating one. Where dispositions to apply terms are neither right nor wrong, dispositions to apply terms given all relevant information seems to enable the discrimination of normative discrepancies between performances. The problem arises when one asks what counts as ‘all relevant information.’ Offloading determinacy onto relevant information simply raises the question of determinacy at the level of relevant information. What constrains ‘relevance’? What about future relevance? Andow chases this inability to delimit complete information to the most extreme case:

It seems pretty likely that there is information out there which would radically restructure the nature of human existence, make us abandon technologies, reconsider our values and place in nature, information that would lead us to restructure the political organization of our species, reconsider national boundaries, and the ‘artificial divisions’ which having distinct languages impose on us. The likely effect of complete information is semantic apocalypse. (Just to be clear—my claim here is not that it is likely we will undergo such a shift. Who is to say what volume of information humankind will become aware of before extinction? Rather, the claim is that the probable result of being exposed to all information which would alter one’s dispositions, i.e., complete information, would involve a radical overhaul in semantic dispositions).

This paragraph is brilliant, especially given the grand way it declares the semantic apocalypse only to parenthetically take it all back! For my money, though, Andow’s throwaway question, “Who is to say what volume of information humankind will become aware of before extinction?” is far and away the most pressing one. But then I see these issues in light of a far different theory of meaning.

What is the information threshold of semantic apocalypse?

Dispositionalism entails the possibility of semantic apocalypse to the degree the tendencies of biological systems are ecologically dependent, and so susceptible to gradual or catastrophic change. This draws out the importance of the semantic apocalypse as distinct from other forms of global catastrophe. A zombie apocalypse, for instance, might also count as a semantic apocalypse, but only if our dispositions to apply terms were radically transformed. It’s possible, in other words, to suffer a zombie apocalypse without suffering a semantic apocalypse. The physical systems underwriting meaning are not the same as the physical systems underwriting modern civilization. So long as some few of us linger, meaning lingers.

Meaning, in other words, can survive radical ecological destruction. (This is one of the reasons we remain, despite all our sophistication, largely blind to the issue of cognitive ecology: so far it’s been with us through thick and thin). The advantage of dispositionalist approaches, Andow thinks, lies in the way it anchors meaning in our nature. One may dispute how ‘meanings’ find themselves articulated in intentional cognition more generally, while agreeing that intentional cognition is biological; a suite of sensitivities attuned to very specific sets of cues, leveraging reliable predictions. One can be agnostic on the ontological status of ‘meaning’ in other words, and still agree that meaning talk turns on intentional cognition, which turns on heuristic capacities whose development we can track through childhood. So long as a catastrophe leaves those cues and their predictive power intact, it will not precipitate a semantic apocalypse.

So the question of the threshold of the semantic apocalypse becomes the question of the stability of a certain biological system of specialized sensitivities and correlations. Whatever collapses this system engenders the semantic apocalypse (which for Andow means the global indeterminacy of meanings, and for me the global unreliability of intentional cognition more generally). The thing to note here, however, is the ease with which such systems do collapse once the correlations between sensitivities and outcomes cease to become reliable. Meaning talk, in other words, is ecological, which is to say it requires its environments be a certain way to discharge ancestral functions.

Suddenly the summary dismissal of the genuine possibility of a semantic apocalypse becomes ill-advised. Ecologies can collapse in a wide variety of ways. The form any such collapse takes turns on the ‘pollutants’ and the systems involved. We have no assurance that human cognitive ecology is robust in all respects. Meaning may be able to survive a zombie apocalypse, but as an ecological artifact, it is bound to be vulnerable somehow.

That vulnerability, on my account, is cognitive technology. We see animals in charcoal across cave walls so easily because our visual systems leap to conclusions on the basis of so little information. The problem is that ‘so little information’ also means so easily reproduced. The world is presently engaged in a mammoth industrial research program bent on hacking every cue-based cognitive reflex we possess. More and more, the systems we evolved to solve our fellow human travellers will be contending with artificial intelligences dedicated to commercial exploitation. ‘Deep information,’ meanwhile, is already swamping the legal system, even further problematizing the folk conceptual (shallow information) staples that ground the system’s self-understanding. Creeping medicalization continues unabated, slowly scaling back warrant for things like character judgment in countless different professional contexts. The list goes on.

The semantic apocalypse isn’t simply possible: it’s happening.