Thought for the Evening: Basic Kinds of Evidence
Are there basic kinds of evidence? One way you might try to handle this is to look at evidence of how people handle evidence. There are three that are particularly interesting.
(1) Pramanas. The root meaning of 'pramana' is measurement, and a pramana is, roughly, a means for acquiring thinking that is correct. In traditional Indian epistemology, the pramana is what bridges the knower (pramatr) and the knowable (prameya). All the major Indian schools of philosophy have some sort of account of pramanas. There are different kinds of pramana; the ones usually you found are perception (pratyaksha), inference (anumana), comparison (upamana), postulation (arthapatti), nonperception (anupalabdhi), testimony (shabda or smrti). You'll sometimes find divisions, e.g., dividing testimony into revelation (smrti) and expert testimony (aitihya). More often, you'll find people reducing the list, by taking some of these pramanas to reduce to others. Since pramanas clearly are evidential in character and an extensive debate exists in the history of Indian philosophy about whether they are basic or reducible, this is obviously an interesting place to look for clues about basic kinds of evidence.
There's a standard account of the distribution of how these were accepted. All the major Indian schools of philosophy, and quite a few minor ones, accept perception (which can, depending on the school, include both sensory perception and internal perception of thought, and sometimes can also include unusual or preternatural perceptions) as a pramana. Only the very strictest materialist schools accepted only perception. Everyone else also accepts inference, in which a proposed conclusion is made certain by a reason tied examples, as when fire is inferred from smoke because we find fire consistently associated with smoke in kitchens and the like. The early Vaiseshika school is said to have only accepted perception and inference. Buddhists do the same. Other schools include testimony (which can include both Vedic revelation and informed human testimony). Samkhya schools, and some others, accept only these three. Everyone else accepts comparison. Comparison, like testimony, is word-based, but it's less direct. If you have never seen a cow, I might describe it; that is testimony. But then you might see the cow, and on the basis of my description say, "Ah, this is a cow!" That's what's meant by comparison as a pramana. The Nyaya schools accept only these four. Everyone else accepts postulation. Postulation is like inference, but less direct. Suppose the sky is very cloudy and it is very dark because of it. I might look at my watch and see that it is past when the sun would normally go down, and say, "The sun must have already gone down." In a sense, I'm profile-fitting; I've learned from other means that there is a normal way things go, and when I have partial information, I fit the normal template to it to conclude the rest. The Prabhakara Mimamsa schools only accept these five. Everyone else also accepts nonperception. If I look at a room and see no jar, I learn there is no jar. Advaita and Bhatta Mimamsa schools accept nonperception. Occasionally you'll find discussion of other groups of people -- calling them schools would be very loose, since it's groups like 'dramatists' -- to which acceptance of other pramanas (like 'fables') is attributed, but it is probably best to read this remarks not as saying that they literally accept these as pramanas but rather that they are accused of treating such things as if they were pramanas.
Such is the standard account, more or less. It actually gets fuzzy around the edges and in the margins. And the differences are not always significant in particular cases -- for instance, the Navya-Nyaya position on postulation and nonperception seems to be not so much that arguments that use them are necessarily wrong but that, when correct, it's because they actually reduce to combinations of the four pramanas that they accept. So sometimes it's an argument over the best way to taxonomize the same things. But, of course, the kinds are what interest us. But, overwhelmingly, the big schools accept perception, inference, and testimony. With the others of the six, the question is usually just whether they are combinations of these three or something distinct.
(2) First principles of common sense. In Beattie's Essay on Truth, a major work presenting the essential ideas of the Scottish Common Sense school of philosophy, a major argument is that all knowledge and inquiry is based on first principles, and Beattie gives us a taxonomy of these. He divides them between those that concern abstract ideas and those that concern really existing things. When talking about abstract ideas, we are dealing with mathematical evidence, which Beattie divides into immediate intuitive evidence and strict demonstration. When talking about really existing things, we are either dealing with our own experiences or other people's. If we are dealing with our own, we are concerned with certainties or probabilities. In dealing with certainties, we are concerned with external sensation, internal sensation, memory, or inference from effects to causes. In dealing with probabilities, we understanding something by using something like it as a model, which can either be the same kind of thing, in which case we are reasoning experimentally, or merely similar to it, in which we are reasoning analogically. And if we are getting our understanding from other people's experiences, that is testimony.
This is obviously divided very differently, but notice that we have perception (mathematical, external sensation, internal sensation, memory), inference (mathematical demonstration, causal inference, experimental inference, analogical inference), and testimony.
(3) Evidential markers. There are languages that have a grammatical feature known as evidentiality. Pretty much all languages have some way of noting the source of one's knowledge of something; but in English, for instance, we mix up evidentiality with modality -- we use modal verbs to handle these things. There are languages, however, that do not do this; they have specific form of grammatical evidentiality that must be used in speaking the language. They are scattered all over, including Native American languages (Pawnee, Western Apache), Brazilian languages, Guinean languages, Siberian languages. These differ quite widely, but there are common patterns. Some evidential markers in languages distinguish between claims describing what you yourself have witnessed and those that you did not witness; others distinguish between claims that are first-hand knowledge of some kind and claims that are not; others distinguish between claims that are reported and claims that are not. When we have more developed cases, we often get the following:
visual witnessed
nonvisual witnessed
inferred
reported
Others that you sometimes find are markers that indicate that something is assumed, or based on direct active participation. Sometimes you get a finer-grain of division, with different kinds of inference or witness or report. For instance, you find languages where claims that are hearsay have to be marked differently from claims that are direct quotations.
There are lots of complications with trying to make sense of how to understand grammatical evidentiality works. My own thought, which is only that, is that we should understand evidentiality as part of the intrinsic etiquette of the language -- as I've noted before, although it's often overlooked, etiquette plays a major and integral role in how languages actually work -- and if that's the case, it's analogous to cases in which you have to adjust your speaking to recognize the particular etiquette-situation. Regardless, what's notable, is that while this is a messy area of language, with lots of variation, the common recurrences are witness, conclusion, report, or, as we could also put it, perception, inference, testimony.
Taking all three of these, we can draw a modest but important conclusion. It's very common for modern theories of evidence to treat all evidence as if it were the same. Bayesian epistemologists are particularly egregious offenders in this regard. However, it seems to be a recurring thing among human beings to handle evidence by breaking it up in to categories. And, while the borders of the categories are often in dispute, the most commonly recurring are perception, inference, and testimony, which are handled distinctly. Now, it's possible that this is just a practical convenience, but unless we have actually shown this, I think it's reasonable to say that we should generally proceed on the assumption that perceptual, inferential, and testimonial evidence are importantly different and work in different ways.
Various Links of Interest
* Mitia Rioux-Beaulne, Fontenelle, Malebranche, et les limites de la philosophie (PDF)
* Bronwyn Finnigan, Phronesis in Aristotle: Reconciling Deliberation with Spontaneity (PDF)
* Zohar Atkins, Beauty Is Peace by Other Means
* The Comedy Wildlife Photography Awards for 2021
* Preston Stuvall, Abductive Inference, Autonomy, and the Faith of Abraham (PDF)
* Joshua Cockayne, We Believe: Group Belief and the Liturgical Use of Creeds (PDF)
* Liam Bright, The Anglo-American Analytic Philosophy Left
* David Kamp, How Jeremy Irons Rescued and Restored a 15th-Century Irish Castle, at Vanity Fair
* Philippe Lemoine, Have we been thinking about the pandemic wrong?
* Ronja Hildebrandt, What Is Philosophy in the Protrepticus?
* Hannah H. Kim & John Gibson, Lyric Expression (PDF)
* Lutherans are on the front lines of the battle for religious liberty
* Stephen Menn, al-Farabi's Metaphysics, at the SEP
* Kevin N. Cawley, Korean Confucianism, at the SEP, which has a really interesting discussion of the interaction between Catholic philosophers and Korean philosophers in the early modern period
* Christian McNamara, The Hidden Life of Ignatius J. Reilly, at "Front Porch Republic"
* Chad Pecknold, Imago Dei as a Political Concept
* Joseph Heath, Why are Racial Problems in the United States So Intractable?, at the American Affairs Journal
* Geoff Shullenberger, Foucault in the Panopticon
* Justin E. H. Smith, Nature Is Becoming a Person, discusses the trend of recognizing various natural features like rivers as juridical persons for certain purposes.
* If it seems to you that movie dialogue has grown harder to understand, you are not alone, nor are you merely getting old and hard of hearing. Here is a discussion of some of the things that have contributed to this problem.
Currently Reading
Matteo Ricci, The True Meaning of the Lord of Heaven
