Accessing knowledge through past theories
You raise the question whether using earlier theories, which are superseded by newer theories, one can (or even could ever) access knowledge? (I am not a physicist, but I do not think that physicists have junked Newtonian theories. I think that Newtonian physics retains its validity in the macroworld we live in and is not considered false. As far as I remember, I have not come across writings that claim that Newton’s theory is false but that most of its consequences (but not all) are derivable in Einsteinian theory, under some ‘limiting conditions’ (like mass remaining invariant, etc.) Anyway, I do not know enough Physics to enter a discussion about it. Arun and Kannan are more competent here than I am.) If we do not get stuck on any one example, it appears to me that there are three questions here: (a) Could defective, incomplete, partial theories help us access knowledge?; (b) Could false, or partially false and partially true, theories help us access knowledge?; (c) Was knowledge ever accessed by past theories, which, as we know today, expressed ignorance?
One route (among many others) to tackle these questions is to treat them as separate (but dependent) questions; the other is to see them as different and independent questions belonging to different domains (epistemology, ontology, etc.); yet another route is setup a framework which sidesteps some unanswered problems that confront us in both these routes. I tend to take the third route. (Ashwin seems to prefer a route that I do not recognize, if I understand his posts correctly.) Which are those unanswerable problems I want to avoid? They are connected to the use of the predicates ‘true’ and ‘false’. Let me very, very briefly explain.
- We have some understanding of what ‘truth’ and ‘falsity’ mean with respect to individual sentences in an interpreted language. Some philosophers have (rightly or wrongly) tried to extend this notion of truth (of sentences) to natural languages. Without seeking controversy here, let us assume that truth is a linguistic property of individual sentences (or the propositions they express).
- Theories are higher linguistic units. (Higher than individual sentences.) Even though we speak of true or false theories (I have done it very often and continue to do so), this talk creates huge issues. The problem is: are theories true and false the way individual sentences are? They could be if theories were ordered sets of sentences. (I have often used this formulation.) However, the only attempt to conceptualize theories as ordered sets of sentences that I know is by Stegmüller and it failed. The current mathematical set theories do not appear to work for theories. (I am talking about scientific theories, of course). So, the question remains unanswered: what kind of linguistic entities are theories?
- How, then, to speak of truth and falsity of theories? One approach, which has also failed, is to develop an ‘inductive logic’, a ‘logic of confirmation’, an understanding of ‘evidence’ etc. Some people are still trying to use and modify Bayesian probabilities to develop degrees of confirmation, robustness of evidence, etc. These are some issues if you want to take the route of ‘truth’ with respect to theories.
- Popper tried the route of falsity. (But because he could not leave ‘truth’ behind, he tried to develop the idea of ‘truth-like-ness’, verisimilitude as it is called.) This attempt has also gone nowhere. The reason for the failure of ‘falsification’ are many: one such is the nature of scientific theories and their predictions. Any hypothesis or theory relies on many ancillary hypotheses and a huge background set of theories and assumptions to make predictions. If any (or even some) turn out to be false we cannot localize falsity: is it the hypothesis, the ancillary theories or the background theories that is false? How to localize it? Called the Duhem-Quine thesis, it said that theories (consisting of the above three) face the tribunal of experience collectively. Thus, the idea of ‘falsity’ of theories confronts other kinds of problems than those that the ‘truth’ of theories confronts.
- Some thinkers took the route of suggesting that individual theories are never tested alone but that rival or competing theories, each with their own history, challenge each other. Such clusters have many names: research programmes, research traditions, etc. Some dropped the predicate ‘truth’ altogether while speaking about the progress of science and scientific theories.
All such discussions were (are) based on serious studies in the history of sciences. They indicate the kind of issues you need to tackle if you tackle the questions you have raised in your mail. I cannot. Therefore, I have taken a different route that sidesteps some of these problems. I am in the process of dropping the notion of ‘truth’ and take the route taken by thinkers above in point 5. (In the meantime, I did some research into the history of ‘truth’ to find out why it has become so important in these discussions. This has convinced me that this notion should not be used either with respect to the world, as ‘facts’ that constitute the basic furniture of the world, or in relation to a theory.)
The chapter on bullshit (in ‘cultures differ…’) tells you that my distinction is between truth and knowledge. Presenting truth as knowledge is bullshit, as I see it. My use of ‘truth’ remains restricted to sentences. Knowledge can be expressed in sentences; knowledge can be accessed also through sentences. However, no combination of sentences (that we want to call a ‘theory’) is knowledge. Knowledge is always true (in the sense that it can be expressed only in true sentences) because I cannot make sense of a ‘false knowledge’. Not all true sentences is knowledge but sentences (or theories) are not the only means of accessing knowledge. When I use the words ‘theoretical’ and ‘practical’ knowledge today, I do not any more use them as different kinds of ‘knowledges’; I see them as different modes of accessing knowledge. I have also given up (after a very great deal of resistance) the idea that “science is the best example of knowledge we have”. Science (or scientific theories with their experiments) is not knowledge, but one of the better-known ways of accessing knowledge. Knowledge can be accessed or expressed in theories, actions, practices, etc., but none of these is an example of knowledge.
Because of this step, not only I can answer the question ‘what is knowledge’ but also why after more than 2000 years of investigation, we still do not know what knowledge is. An analogous situation exists with respect to science: histories, philosophies, psychologies, sociologies, anthropologies… of science, including science of science, have failed to tell us what science is. We really do not know what science is today. This failure too requires an explanation.
Thus, to answer your question from where I am currently: partial, defective, incomplete theories give a partial, defective and incomplete access to knowledge. False theories do express ignorance and they could either facilitate or hinder the emergence of knowledge in an individual. If they hinder, probably you are dealing with bullshit (i.e., truth parading as knowledge). On their own, theories are neither the means nor are they hindrances to access knowledge. Only in an individual does knowledge emerge (thus only an individual can access knowledge); but for this to happen that individual must be in a particular state, which is both physical and mental. Put in a banal form: you must undergo preparation if you want to acquire knowledge. (You cannot acquire knowledge of Einsteinian physics, if you have not put in the required effort and preparation.)
The framework and the story that outlines this route is planned as a chapter in the forthcoming book.