Is Knowledge Justified True Belief? - the tripartite theory and Gettier's counter examples
Traditionally knowledge, at least propositional knowledge, has been defined by philosophers as justified true beliefs. In order for beliefs to be regarded knowledge somebody had to believe it, there had to be justification for the belief and the belief had to be true. This though has been contested in recent times; primarily by Gettier and the counter examples he found. Justified true beliefs still seemed to remain a requirement for knowledge but often seems insufficient on its own, something else is needed to make it a reliable definition of knowledge.
The tripartite theory of knowledge as justified true belief has always been seen as necessary conditions for knowledge. The clause of justification has proved almost necessary as a belief which is true without justification seems only to be a belief which is true through shear accident. The requirement that the knowledge has to be believed is obvious in itself as without belief its truth and justification are irrelevant. Truth is of equal importance to both other clauses as without truth again it does not seem to be knowledge at all, it would be a false belief, and you cannot view the world with false ‘knowledge’. However, these conditions, though wholly and inarguably necessary, do not seem sufficient alone.
This customary idea that knowledge of justified true belief has been contested by some, primarily Edmund Gettier. In his paper ‘Is Justified True Belief Knowledge?’ Gettier was able to show that, although these terms are necessary, ‘the definition of knowledge must be modified either by adding a fourth condition or by building an appropriate clause’(Steup, M. 1998; 5). An example of a Gettier counter-example is if someone was to watch what they believed was the live show of a football match of Aberdeen against Celtic and Celtic won 3-1 but in fact they had been watching a recording of a previous match but the results of the actual match were the same would their belief of the score be knowledge? Examples such as these do fit the conditions of causing true and justified beliefs but it is hard to truly consider them knowledge.
A lot of philosophers have attempted to come back at Gettier’s counter examples by trying to show that the counter-examples don’t work. One way to do this is to look at the principles of inference the counter-examples use. ‘It must be possible for a false belief to be justified and a justified belief must justify any belief which it implies’(Dancy, J. 1985; 26). Proving these principles false or unreliable could undermine the counter-examples. However, it could be possible to construct new counter-examples which do not rely on inference so these criticisms have failed to conclusively defend the tripartite theory. It seems that Gettier’s counter examples are perfectly effective in their own terms.
There are also many responses to Gettier which attempt to remedy the problems with the tripartite theory of knowledge itself. The most obvious argument is that the initial belief is false and it is from this that the true and justified belief is inferred. It is easy enough to add a fourth condition which states that ‘nothing can be known which is inferred from a false belief, or from a group of beliefs of which one is false’ (Dancy, J. 1985; 27). However, there are counter-examples which can be written to contain a falsehood but no inference. For example, I believe there is a cat in the house next door because of what I can see through the window. What I am actually seeing is a small dog, but I am not wrong as there is also a cat in the house. There was no inference, I simply take myself as seeing a cat; the belief is true and justified but it’s still hard to think I knew there was a cat. Too argue further that this sensory perception does not qualify as knowing would cause issue as it is likely to make it near impossible to know anything.
Another approach sees the Gettier counter-examples as the result of there being truths which would have caused the believer’s justification to fall short had be believed them. So in the example of someone watching a recording of a football match, if they became aware that it was a recording their belief that it showed the results of the live match would be destroyed. This risks proving the justification clause for knowledge unnecessary; ‘it looks as if a false belief could never be indefeasibly justified’ (Dancy, J. 1985; 29) this is because there would always be a truth, even the negation of the false belief, the addition of which would act to ruin the justification. This, although seeming a weakness, does at least render the truth clause an evidential requirement rather than mere stipulation as it was before.
Reliability is also brought to attention as a possible route of attack in order to defend the tripartite theory from Gettier’s counter-examples. This argues that justified true belief can be regarded knowledge when it is taken from a reliable method. In the example of somebody watching a football match, they had known the match was being played that afternoon, derived from reliable newspapers normally right about this. The belief that the results were what he saw of a recorded game though are less reliable. This method requires an account of what reliability entails. It could mean a method which ‘if properly followed, is perfectly reliable and never leads to a false belief’ (Dancy, J. 1985; 31). This though seems an unlikely method as there doesn’t seem to be many, if any, perfectly reliable methods; apart from anything else, man is fallible and so methods available to man are likely to also be fallible. However, making it a generally reliable clause opens it up to criticism as this requires that methods found unreliable in other situations be regarded reliable enough for this situation. It is possible to further suggest that it is only necessary for the method to be reliable in this particular situation. This though seems a rather indefinite addition to the tripartite theory as the reliably relies either on justification which adds nothing to the truth condition or vice versa.
Another possible failing in the Gettier counter-examples could be that the reasons were not wholly conclusive. If the tripartite theory is based on conclusive reasons then all the Gettier counter-examples fail. One possible account of what it is to be conclusive is ‘where beliefs A-M constitute conclusive reasons for belief N, A-M could not be true if N were false’ (Dancy, J. 1985; 32). Although this does manage to exclude all of Gettier’s counter-examples it also acts to make knowledge a rare possibility, especially empirically. A weaker account is offered by F. Dretske who suggested that these ‘reasons A-M for a belief N are conclusive iff A- M would not be true if N is false’ (Dancy, J. 1985; 33). That A-M would not be true if N were false is not as strong as to say they could not be true but this weaker account does not convey a sense of conclusiveness at all. This though does not mean it doesn’t have some virtuous factors.
A proposal is put forward by A. I. Goldman regarding a causal addition to the tripartite theory. This idea comes from the observation that Gettier’s counter-examples could purely be regarded as luck that they fit the tripartite theory. The condition that luck not be involved though would be impossible to impart as often a sense of luck is relied upon by most people. Goldman suggests that the ‘fact that p should cause a’s belief that p’ (Dancy, J. 1985; 33). So coincidentally true beliefs such as the Gettier counter-examples could be disregarded. This approach though also faces problems as surely facts do not cause anything; they merely reflect the world rather than affecting it. Further, Goldman’s example does not included the possibility of knowledge about he future as this would require backwards causation of the future causing the past. Finally there is a problem of knowledge through inference as this would be impossible to regard as knowledge according to Goldman’s theory. It would be impossible to know, for example, that all mammals die as this is based on the perceived instances of mammals dying and these mammals did not die because all mammals die. It does though seem to be a knowable fact that all mammals die. The first two criticisms do have further criticisms to them making their status as criticism questionable. The third criticism though seems more sturdy.
In conclusion, although the tripartite theory seems to be a requirement for knowledge, alone it does not stand up to criticism. The Gettier counter-examples prove this insufficiency of the tripartite theory of knowledge. Despite the many attempts to respond to Gettier, all have some flaw. Additions and supplements to the tripartite theory have all failed to conclusively further the theories reliability as a method for gaining knowledge, though some seem more successful than others. The question can be raised whether there is really any point in finding a definition of knowledge or is it ‘a mere technical exercise?’(Steup, M. 1985; 26). Would anything rely on the successful definition of knowledge? It seems that there could be a possible use for at definitive definition as it may ‘suffice to undermine crucial sceptical moves’ (Dancy, J. 1985; 27). Thus rendering the finding of a definitive understanding of knowledge an ongoing search which remains, for now, unsuccessful.