Beyond Stagnation – Things to Come

I am currently gathering and editing some posts from late last year into a working paper. The general theme is one I picked up from Tyler Cowen, stagnation, and my working title for the paper is Stagnation and Beyond: “Intangible factors” and the next frontier. Here’s a bit of new material I’ve added to the paper. It comes after a discussion of Patrick Collison and Michael Nielsen, “Science Is Getting Less Bang for Its Buck” (The Atlantic, Nov 16, 2018), where I introduced the concept of cognitive ranks as a way of uniting two metaphors they propose for scientific advance, exploration (1) and the endless frontier (2). Revised 7 July 19.

* * * * *

In scientific prognostication we have a condition analogous to a fact of archery--the farther back you are able to draw your longbow, the farther ahead you can shoot.
– R. Buckminster Fuller, Critical Path

The obvious question is: What next? What’s the new cognitive architecture that will move us beyond the current stagnation? I don’t know. Though I do have some scattered thoughts on the matter.

First, I observe generally, that we should not think that the consolidation of new systems of thought will solve problems we currently find both very important and intractable. Sure, maybe some of them will be solved. But it seems just as likely that some will simply be dropped as no longer being worthy of pursuit. The emergence of an effective chemistry, for example, did not show us how to transmute base metal into gold, the dream of an earlier age. What of the current situation in the foundations of physics [1], where theoretical speculation has outstripped empirical technique for several decades? Will the dream of a grand unified theory go the way of the philosopher's stone? I haven’t a clue. But some current intellectual dreams will end up in the dust bins of history. Just as surely, there will be surprises going forward. Thus, on the whole, I’m sympathetic with Collison and Nielsen when they say:

If we see a slowing today, it is because science has remained too focused on established fields, where it’s becoming ever harder to make progress. We hope the future will see a more rapid proliferation of new fields, giving rise to major new questions. This is an opportunity for science to accelerate.

Second, the study of culture may well come of age over the next several decades, as biology did, say, during the century between Darwin’s On the Origins of Species (1859) and Watson and Crick, “Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic Acid” (1953). Culture is studied in a variety of intellectual disciplines, some humanistic, others more social scientific in orientation – for what it’s worth, I’m partial to “the human sciences” as a term, which is more European than North American in provenance. I’m not imagining that these various disciplines are going to come together in some “Kum ba yah” of methodological and theoretical comity.

I’ve got something messier in mind. In the decades World War II anthropologists and archaeologists developed methods for conducting statistically controlled cross-cultural studies of various kinds, including cultural complexity. More recently we’ve seen a variety of empirical work conducted in various disciplines – anthropology, linguistics, communications, sociology, literary studies, popular culture, and so forth – tracking change in cultural phenomena. Some work, of course, has taken advantage of social media and has tracked propagation of ideas (“memes”) across the web. Other work involves controlled experiments of transmission of “memes” from one person to another (reminiscent of the classic work of Frederic Bartlett from the second quarter of the previous century).

Something is going to come of this, though just what and when I’m unwilling to prophesy. The major problem I see is in the description of cultural phenomena. Here I’m thinking of the centuries of descriptive work that Darwin had at his disposal [2]. Consider the case of linguistics. While the work of Noam Chomsky precipitated a great deal of interest and excitement in the last half of the previous century, we do not see widespread agreement among linguists on the proper way to describe the facts of syntax. Perhaps the most visible manifestation of this situation is the rather nasty quarrel between Chomsky and Daniel Everett, and their partisans, over the role of recursion in language [3]. And if linguists can’t agree on the description of sentences, what hope is there of literary critics agreeing on the description of texts?

Still, I think it reasonable to believe that in the next few decades something will come of this. Whether or not it will be a “science” of culture, who knows? But it will be something. Whatever it is, though, it will not answer all the questions John Horgan raised in The End of Science. It will be a new intellectual world. Just what will become of those questions, I don’t know, though I’m happy to believe that some of them will go the way of the alchemists search for a way of transmuting base metals into gold.

And so we arrive at my third set of observations. Perhaps I can say something useful by using the idea of a coming technological singularity as a conceptual foil. The idea seems to be that some as yet undetermined time in the future computers will exceed the human brain in raw computational capacity. At this point the machines will be able to think just as well as us and, in short order, they will recursively bootstrap themselves into ever higher levels of competence. Just how this will happen seems not very well specified, as though it may well be an act of spontaneous conceptual combustion happening within the machine itself without human intervention.

Here’s what Robin Hanson has to say about Foom, his term for this magical machine florescence:

“Intelligence” just means an ability to do mental/calculation tasks, averaged over many tasks. I’ve always found it plausible that machines will continue to do more kinds of mental tasks better, and eventually be better at pretty much all of them. But what I’ve found it hard to accept is a “local explosion.” This is where a single machine, built by a single project using only a tiny fraction of world resources, goes in a short time (e.g., weeks) from being so weak that it is usually beat by a single human with the usual tools, to so powerful that it easily takes over the entire world.

I would further add that this machine has to build its way through several ranks of human cognitive ontologies, one after the other, just as humans do during the course of their education. As far as I know, for example, no human has ever learned quantum mechanics, e.g., straight out of grade school without first having mastered several “layers” of mathematical and scientific prerequisites. That’s just not how things go.

What Hays and I have been arguing is that, over the long-term of cultural evolution, we manage to recursively bootstrap ourselves into ever more effective families of cognitive architectures, which we call ranks. Rank 1 is grounded in our capacity for speech; Rank 2 is grounded in writing; Rank 3 adds effective arithmetic calculation into the cumulative mix; and Rank 4 is learning to exploit computers and computation [5]:

We are suggesting that these four ages, the systematic differences between cultures at these four levels of cultural evolution, are based on differences in cognitive mechanism. As cultures evolve they differentiate and become more complex and sophisticated, the more sophisticated cultures having cognitive mechanisms unavailable to the less sophisticated. Over the long term this process is discontinuous. That is, at some point in the evolution of a culture a new kind of thinking becomes available which permits a dramatic reworking of culture. This new kind of thinking is engendered by a new capacity for manipulation of abstractions. [...] Each cognitive process is associated with a new conceptual mechanism, which makes the process possible, and a new conceptual medium which allows the mechanism and process to become routine in the culture. This is an important point. The general effectiveness of a culture is not determined by the achievements of a few of its most gifted members. What matters is what a significant, though perhaps small, portion of the population can achieve on a routine basis. The conceptual medium allows for the creation of an educational regime through which a significant portion of the population can learn effectively to exploit the cognitive process, can learn to learn in a new way.

However skeptical I am about machine superintelligence I have no doubt that – barring a civilization-threatening catastrophe such as nuclear war, environmental collapse, or both – a new level of thought will precipitate out of our ongoing – stagnant? – and ever changing flux. Computational tools and systems of all kinds will be central to this consolidation. We will find ourselves looking ahead to a brave new world.

What then of the endless frontier? For any given rank of cognitive architectures there will be a limit, a point in which the territory has been more or less thoroughly explored. But, as far as I can tell, there will always be the possibility, the task, of evolving a new rank of architectures. That new rank presents us with a new territory to be explored, and new frontier. We enter it, explore it, exhaust it.

Consider the previous curve with a different set of labels:

Is that what’s going on? Does stagnation mark the shoulder of a growth curve, the phase where exploration is exhausting what was once a new frontier while the cultural-cognitive system is preparing to reorganize and consolidate around a new underlying cognitive architecture? As that new architecture consolidates what do we see before us but a new frontier?

That succession of phases (explore → stagnate → consolidate) gives us, in effect, an endless frontier. To understand how it works, to explore it, we must take explicit account of both the structure of the world and the structure of thought.

Of course, that cultural-cognitive system is us. We need to recognize what is happening and take appropriate measures.

Do I believe that? How do I know? I just made it up, but it’s my best current speculation.

References

[1] Sabine Hossenfelder, Lost in Math: How Beauty Leads Physics Astray, Basic Books, 2018.

[2] See, for example, Brian Ogilvie, The Science of Describing: Natural History in Renaissance Europe, University of Chicago Press, 2006.

[3] This dispute is at the center of Tom Wolfe’s final book, The Kingdom of Speech, 2016.

[4] Robin Hanson, “I Still Don’t Get Foom”. H+ Magazine, July 29, 2014, https://hplusmagazine.com/2014/07/29/i-still-dont-get-foom/.

[5] William Benzon and David Hays, “The Evolution of Cognition”, Journal of Social and Behavioral Structures 13(4), 297-320, 1990, https://www.academia.edu/243486/The_Evolution_of_Cognition.