Rauch on Peirce (Wrong 3)
a priori vs. scientific methods in Fallibilism
In my first post about Fallibilism I mentioned that Charles Sanders Peirce didn’t provide a lot of detail about the Method of Scientific Investigation, which is the fourth and best method of fixing belief. In my second post I devoted a bit of space to discussing what Peirce called the A Priori Method, which is the third method on his list. As Peirce makes very plain, it’s also the most slippery and deceptive, in that it seems reasonable on its face but can spackle over a lot of faulty logic and questionable assumptions.
Jonathan Rauch, who first made me aware of Peirce’s philosophy in his book The Constitution of Knowledge, has read more deeply into this topic than I have. He dropped me a line the other day to provide some clarification of how the A Priori Method differs from the Method of Scientific Investigation. As a reminder, these two methods are superficially the same in that both of them eschew the methods of Tenacity and Authority, which are obviously irrational. Both A Priori and Scientific Investigation purport to base beliefs upon rational analysis of observable facts. What then is the difference between them?
Here’s a quote from Jonathan Rauch:
I thought I'd address a question you allude to re the difference between the "a priori" method of settling opinion vs science. Here's how I understand it...
A priori methods are superficially similar to science, in that they posit hypotheses and use evidence and rational argument. But they're still fundamentally individualistic. Individuals, or factions, form conclusions and argue with each other, but even if they're brilliant, they have no systematic way to bring their views toward consilience (a collective, multi-perspective resolution). Whereas science is inherently a collective endeavor in which, as Peirce says, only "we" matter, not you or I. Knowledge arises as an emergent phenomenon from the comparing and collating of multiple points of view. Though he didn't have the term, Peirce invented network epistemology. Reality lives on the network, a property of the hive mind.
My favorite example: if you see a disheveled man scribbling equations in a garret apartment in Bern, you have no way to know, even in principle, whether he's Einstein or a madman. Science only begins when his claims enter the network and undergo collective evaluation.
It's a sophisticated concept...and a human breakthrough.
Since this helped clarify the distinction for me, I thought I’d turn it into a quick post for anyone who has been following my occasional notes about Fallibilism.
Rauch (and Peirce) are correct about the importance of collective or consillient knowledge. But it goes even further than that. The network of knowledge must extend not only among the collective minds of society, but also among the collective knowledge of society. There is a very real way that we can say that the Gorilla was discovered in Africa in 1847 even though it had been known about by humans for millennia. Because while it was known, the knowledge about them had not been integrated into the rest of our collective knowledge. When it was described "scientificially" that meant primarily that the facts about it were inserted into the rest of all we knew in a way that was consistent and coherent. There are pools of knowledge that may be "true" — such as Shamanism, mysticism, psychism — and that are even coherent within themselves — that we cannot integrate into everything else we know about the world, so they are excluded from science. If we are able to integrate what is new into everything else we know, we call that collective or network of knowledge, science. I like to think of science as the way we know things. Besides the other aspects of how we know things via science, such as being able to falsify it, the ability to integrate it into everything else we know is vital. Science is indeed a network, but also a network of facts.
I think we are likely to eventually conclude that intelligence is inherently recursive, which makes the distinction between individuals and collectives somewhat vacuous. Individual minds ARE networks.
It’s still possible to have conscilience in an individual mind - you can run the Hegelian dialectic on your own if you want. Conversely, it’s possible for a hive mind to ape at “debate” while rejecting evidence that doesn’t fit its conclusion.
So rather than ignoring individual human beings and saying only the network matters, I think the true distinction that matters is:
- the evidence available on the network
- the processing power of the network
- the reward function allocating resources on the network
The last one is critical. Large groups often get things wrong that individuals can see correctly, because the groups are driven by reward functions that prioritize status within the group now, over advancing the group’s long term survival.