 | |
| From http://superscholar.org/interviews/christopher-michael-langan/ |
Someone pointed me recently to Christopher Michael Langan’s long paper on his CTMU -- Cognitive-Theoretic Model of the Universe. I took a little time today to read it over, and here are some of my thoughts....
This “new model of the universe” has received little if any scrutiny in the academic literature, but it has gotten some media attention in the past, due to its author being an interesting character— an especially high-IQ individual who is also a body-builder and has worked throughout his life in blue collar jobs rather than becoming an academic or a corporate researcher…. (Outliers of this sort tend to interest me personally, as I was a semi-prodigy myself in youth, generally waaaaaay ahead of the curriculum while going through school, and finishing my undergrad degree at 18 and my PhD at 22 ... and before finally deciding to finish my PhD and become an "official researcher" I often considered taking a pure-maverick approach similar to Langan's....)
In capsule summary, Langan views reality as a language for talking about itself, to itself. He understands time as an emergent relationship between different languages, “later” ones along some “emergent timeline” extending previous ones. He looks at series of languages as progressing toward goals that are expressed as maximization of “generalized utility functions.” In this view, mind and physical reality are seen as part of the same set of networks of linguistic relationships. Individual minds are connected with the universal mind, and with physical entities, via linguistic relationships.
He expresses these interesting ideas (and many more details, and some points my crude summary above skips) using a lot of idiosyncratic terminology, plus some reasonable mathematics…
One term he uses for his model of reality is: a ”Self-Configuring Self-Processing Language" or SCSPL
The only thing in his model that really rubs me the wrong way is his placing a “utility function” in a central role. This seems a limiting perspective, as Weaver and Viktoras pointed out in their recent paper on “Open-Ended Intelligence.” But in this regard, Langan is more mainstream than me; the reinforcement-learning perspective is very popular in AI right now, and optimality principles have been central to physics for centuries…
Other than that, the overall framework he posits makes sense to me.
As Langan himself affirms, what he’s putting forth is a philosophy, not a scientific theory. It’s not really a new kind of “theory of everything”, it’s a mathematically semi-formalized metaphysics…. If his theory hasn’t attracted much attention yet, I think this is not just because of its author's "outsider" status, but also because almost nobody cares that much about formal metaphysics these days…. Current culture focuses mainly on practical things; and those concerned with metaphysics are generally new-agey and experientially-oriented, not formalism-oriented.
A few further detailed comments follow below…. These are basically my own notes upon reading Langan's paper, rather than a structured formal critique or response, so take them for what they are...
Some More Detailed Comments
Comment 1)
Most of what occurs in Langan’s theory has been said before, though with different terminology and sometimes in different forms.
The idea that the universe can be formulated as a language that is talking to itself, about itself, has been said many times before; and I wrote about this myself in my book “Chaotic Logic” and later in “The Hidden Pattern” (including references to others discussing these concepts).
The “telic recursion” aspect he discusses is not laid out in enough detail to really understand what he means.
“ Telic recursion occurs in two stages, one primary and global, the other secondary and local. The primary stage creates the distributed laws, including the laws of physics, which reality obeys, while the secondary stage creates nondistributed, ad hoc supplements to those laws as reality transitions from state to state”
This seems very closely related to Peirce’s “tendency to take habits.” That is, if the universe at a certain moment (in a proto-temporal sense, not necessarily the same as the time-axis of physics) manifests certain patterns, then fulfilling these patterns in the “future” of that moment can be viewed as a “goal” which is being achieved via following the “tendency to take habits.” But this assumes Langan’s “distributed laws” can be equated to Peirce’s “habits”, which isn’t entirely clear because Langan’s idea of telic recursion isn’t spelled out in enough detail to fully understand it.
Going on in this vein, if you put together the “cognitive equation” from my book Chaotic Logic, with Peirce’s “tendency to take habits”, you get something pretty close to Langan's telic recursion. And my formalization of infinite-order probability theory is an alternative, and potential complement, to his language-theory formulation.
I’m not dissing Langan for not referencing my own works, which are somewhat obscure and I wouldn’t expect him to know. But I’ll note that when I developed these ideas, I gave lots of references to, and discussions of, prior thinkers with similar ideas. Langan is under no obligation to do this, and his own style has its advantages. But it also gives the casual reader an impression that these are wildly new ideas, when in fact most of this has been said before in various ways and context.
OTOH, as Pascal said (something like), “Let no one say there is nothing new. The arrangement is new.” Arranging familiar ideas in a new way can have significant value, and it may be that Langan’s arrangement of these ideas -- together with some novel bits and pieces -- has some particular force. I’m not decided on that yet.
Comment 2)
Metaphysically, Langan’s theory seems restricted to Peircean Thirds.
Langan says “everything essential to reality, including everything needed to describe it, is contained in reality itself”. But this assumes “everything” is contained within the realm of description — i.e. it rules out Peircean Firsts and Seconds. This is implicit in the “everything is language” approach. Langan may well see this as a feature; I see it as a bug….
Comment 3)
Langan mentions John Wheeler’s desire to found physics on “It from Bit” and observation and related ideas. Wheeler's ideas are indeed fantastically inspirational; and Langan wonders why nobody has done what Wheeler envisioned. But actually, Langan has not done so either (at least not yet!). Wheeler wanted to actually derive physics from these philosophico-mathematical foundations, not just formalize them as a mathematical metaphysics.
Regarding physics, the connection of Langan’s theory with quantum theory is left pretty durned vague. There is a lot of work in the last couple decades deriving quantum theory from various very abstract axiom sets (e.g. the relational interpretation). Can he do something similar or related here?
The connection w/ physics theories of spacetime is left very vague as well. Is there some connection e.g. with causal set theory? Causal set theory lacks a propagator (so you can’t use it to derive QFT) … does his metaphysics place some constraint on the propagator that reality must use?
If there are connections of this nature between his metaphysics, and physics, that would be interesting…. Various others are engaged with attempting to derive quantum theory, essentially, from metaphysics. They haven’t quite succeeded but have “almost succeeded” in various interesting ways… I hope this is where he aims to take his theory next….
Comment 4)
There is a statement on the CTMU wiki to the effect that Langan’s theory proves the existence of God. Elsewhere Langan has talked about patching up holes in Anselm of Canterbury's 11th century proof of God's existence.
I'm more impressed with CTMU as a theory of reality, than as a proof that God exists. In this vein, to put it a bit too crudely, what Langan shows in the CTMU paper so far is something like
I'm more impressed with CTMU as a theory of reality, than as a proof that God exists. In this vein, to put it a bit too crudely, what Langan shows in the CTMU paper so far is something like
- reality is omniscient, in that everything in reality is known to reality
- reality is omnipotent, in that reality is able to do everything that is possible to do
- reality is omnipresent, because it pervades all reality
But this sort of argument doesn’t really justify anything resembling conventional (e.g. anthropomorphic) religious “Gods”, nor does it argue in any useful sense that the universe was designed by something carrying out reasoning or deliberative processes…. The connection between Langan’s ideas and “intelligent design” seems a red herring to me. His theory suggests that the universe as a whole may be thought of as a mind or intelligence, but doesn’t say anything about this universal mind carrying out a design process (as opposed to, say, a more evolutionary sort of cognitive process).
Medieval-style arguments for the existence of “God” don’t interest me much, but fortunately Langan’s investigations have many more interesting aspects…-
On Academia Etc.
Finally, a few comments regarding Langan’s choice to pursue his theorizing purely as a maverick individual thinker, rather than within the academic world
This choice is very understandable to me —- certainly I find that academia can be tedious and stultifying, with its emotionally narrow culture and its focus on incremental progress, on boring false-modest writing, and on paying service to the recognized lions in each field of study....
On the other hand, one thing academia DOES have is a time-honed set of mechanisms for propagating ideas beyond the individual to the community. And Langan does seem to be missing this -- his maverick ideas are deep and intriguing, but they don't seem to be getting followed up by a community of others. Which is too bad, because they're interesting and could likely benefit from the sharpening and expansion that a community of minds focusing together can provide....
I find myself mentally comparing Langan to Eliezer Yudkowsky, another high-IQ maverick who has personally avoided the academic establishment, while developing his own deep and idiosyncratic view of the universe. Both Langan and Yudkowsky have the habit of introducing a lot of novel vocabulary for describing their ideas, though they have different styles of doing so (Langan likes inventing new words; Yudkowsky prefers assigning new meanings to commonplace phrases, e.g. “Friendly AI” or any of the zillion other “defined terms” commonplace on the Less Wrong blog/network he founded).
So far, it seems Eliezer has achieved his greatest influence via passing ideas along to more conservative academic types like Nick Bostrom (whose recent, celebrated tome Superintelligence seems to consist largely of ideas borrowed from Yudkowsky and then further polished), and the more conventional mathematicians currently doing math theory at MIRI (the organization Eliezer founded, as SIAI).
So one might say Yudkowsky has achieved his impact via playing on the fringes of academia and inviting more conventional academics to help develop his ideas — whereas Langan, whose ideas are also interesting, has played a bit further away from the edges of conventional academe, and hasn’t yet attracted so many followers. But of course their distance from the edges of academe isn’t the only difference between the two — e.g. another difference is that Yudkowsky is damn good at self-promotion; and another one is that Yudkowsky’s area of obsession, AI, happens to have become very popular lately ... way more so than metaphysics and formal theology....
But anyway…. These issues of science-community politics interest me personally because I’ve spent my own career playing around the fringes of academia in various ways (though unlike Langan and Yudkowsky I went a more conventional route and got a PhD, and then spent the first 8 years of my career as a professor before deciding that the incrementalist and status-hierarchy-focused style of academia wasn’t necessarily my cup of tea)…. But these sociological points aren’t ultimately as interesting as modeling the nature of reality! Regarding which, Langan’s ideas are well worth reading and thinking about. They don’t yet lead anywhere dramatically conclusive or pragmatically impactful (IMO), but they plant interesting ideas in one’s mind, which is pretty much what philosophy is supposed to do…. I'm curious for what comes next....