Copyright © 2013 Bruce W. Hauptli

Chapter 6. Coherence and Observation:

This Chapter discusses and develops a coherentistic conception of observational beliefs.  It is best to begin this chapter by reading the final summary paragraph on p. 138--it helps clarify what is to come.  Observational beliefs will play a central role in BonJour's overall theory, and in his argument to the effect that the coherentistic theory which he develops overcomes the traditional objections to such theories (and resolves the "regress" problem of justification).  Since these observational beliefs play a central evidential role, their justificational or evidentiary status is clearly one thing which may be questioned--one may be tempted to question whether his appeal to these beliefs (or a particular coherentist's appeal to such beliefs as she justifies some other beliefs) is legitimate--in other words, does the postulation of a specific sub-set of beliefs constitute an acceptance of the thesis of "epistemic priority," and does it engender a skeptical challenge?

     The detailed argumentation of this Chapter is hard to understand unless it is placed in this sort of context.  Suppose I ask you if you have a quarter and you reply "Yes."  How will you respond to a justificatory query here--to the question "How do you justify your belief that you have a quarter?"  Now suppose I ask you if you have a belief regarding the color of my hair (or the color of this piece of paper), and you reply "Yes."  How will you respond to a justificatory query in this case?  Before responding, take special note that the justificatory query here is not about the color of my hair (or of the piece of paper).  It is a query as to how you can justify your particular belief.  How do you justify your observational beliefs?  Remember also, that BonJour must offer a coherentistic justification here--if his treatment of observational beliefs assigns them a foundational status, he will be a foundationalist rather than a coherentist.  On the other hand, of course, if his observational beliefs are just like all the other beliefs in his system, then they will not be able to perform the function he assigns to them.

6.1 An Initial Objection:

112 If we are to meet the "input objection," we must distinguish between how a belief is arrived at and how a belief is warranted or justified.

-113 "...a belief may be said to be inferential or noninferential in two quite different senses....observational beliefs are obviously and paradigmatically noninferential in the first sense...they are noninferential in origin....But it is not immediately obvious that a belief which is thus noninferentially arrived at must also be noninferential in the second sense...[that it] must possess noninferential warrant."

6.2 A Suggestion from Sellars:

115 Wilfred Sellars suggests that the initial credibility of an observational belief may be a function of how it comes to be accepted:1

-[Sellars] "...certain empirical belief tokens2 possess a credibility or warrant which does not depend solely on their content, but which is instead somehow a function of the way in which each particular belief token comes to be accepted, that is, as it was put earlier, its origin or genesis in a person's thinking."

-116 [BonJour] "...Sellars characterizes the process which culminates in...an observational belief (or statement) as a "language-entry transition."  In a language-entry transition a person comes to occupy a "position" in the linguistic or conceptual "game" by virtue of a stimulus-response connection in which the stimulus is not a position in the game but the response is such a position.  The beliefs which result from such a process are caused from outside the system of beliefs but justified only from within the system; and moreover their justification depends on their having been thus caused."

116-117 According to BonJour, "something like Sellars' view of observation seems to me the basic ingredient which is required if a coherence theory of empirical knowledge is to be even a prima facie candidate for a correct account of empirical knowledge."

6.3 Coherentist Observation: An Example:

117  "...I do not believe simply that there is a red book on the desk, but rather that there is a book of a certain approximate size, of an approximately rectangular shape, which is a certain fairly specific shade of red, and so on...."  This belief is cognitively spontaneous--I do not infer that there is a book on the desk.

-117-118 it is a belief of a certain specifiable kind K₁ (e.g., visual belief about the color of a medium-sized physical object);

-118 it is arrived at under specifiable conditions C₁; and

-for a class of individuals, this sort of cognitively spontaneous belief is highly reliable.

Where these conditions apply, the following justificatory argument for the cognitively spontaneous observational belief may be offered:

(1)  I have a cognitively spontaneous [visual] belief of kind K₁ that there is a red book on the desk.

(2)  Conditions C₁ obtain.

(3)  Cognitively spontaneous visual beliefs of kind K₁ in conditions C₁ are very likely to be true.

[4]  Therefore, my belief that there is a red book on the desk is very likely to be true.

[5]  Therefore, (probably) there is a red book on the desk.

--119 Criticism: Before discussing these cases, BonJour says that he will "...not trouble to distinguish between the actual facts of each situation (regarding whether the belief is cognitively spontaneous, whether the conditions C₁ obtain, and whether such beliefs are likely to be true) and my subjective conception thereof, but will simply assume that the latter is in accord with the former except where specified to the contrary; allowing for the opposite possibility would greatly complicate the discussion, but would not significantly affect the main issue."  However, the skeptical possibility that our experiences and our beliefs about them might diverge is a consideration he holds against the foundationalists, and it is certainly a significant opening for the skeptic (whose orientation BonJour must address if his overall argument is to be successful).

--119-120 Examples of "contrasting cases" where such an argument is not available although we have cognitively spontaneous beliefs: the belief that the car which just went by is a Lotus (not a belief involving an ordinary medium-sized physical object [the "kind" is too specialized]); the belief that someone far off in the distance is a particular friend [conditions C₁ don't obtain]; believing that one sees someone specific in the dark [not likely to be true in these conditions]; and seeing a fat man in a hall of mirrors [conditions C₁ don't obtain].

121-122 Discussion of negative observational knowledge [that there is no blue book on the desk] and contrast between it and positive observational knowledge.  Note that BonJour is not claiming that we "observe" negative things but, rather, that such knowledge "depends closely on observation."

122 BonJour maintains that his "...observational beliefs are epistemically justified or warranted only in virtue of background empirical knowledge which tells me that cognitively spontaneous beliefs of that specific sort are epistemically reliable...under the conditions then satisfied.  This in turn suggests a general account of observation, at least broadly coherentist in character, involving the following essential conditions":

-"There must be a class or category of cognitively spontaneous beliefs which are distinguishable and recognizable by the person who has them."  [K₁]

-123 "The class of cognitively spontaneous beliefs in question must be epistemically reliable in relation to the subject matter of those beliefs...."

--Problem: BonJour recognizes that the observer "...has no epistemologically unproblematic access to such objective reliability...[thus, reliability must be] judged from within the person's system of beliefs...."  But doesn't this just simply reintroduce the "input objection?"  That is, the notion of observation here is supposed to provide a connection with the "external" world, but reliability is referenced back to the coherent system rather than to the world!  BonJour responds that this will be addressed as he considers the problem of truth later.

-"The believer in question must himself have cognitive access to the required justificatory premises: He must be able to recognize beliefs of the kind in question and distinguish them from others....He must believe that beliefs of this kind are...reliable....it must be possible for him to believe with justification in a particular case that the requisite conditions for reliability are indeed satisfied."

--Note: taken too explicitly, this will preclude visual knowledge in children and those who haven't taken the requisite physiology courses!  Taken too loosely, it could be empty of content!

6.4 The Justification of The Premises [of the five-step justificatory argument for the cognitively spontaneous observational beliefs]:

  124 Justification of premise (3): that cognitively spontaneous visual beliefs of kind K in conditions C are very likely to be true:

-This is justified in the same [coherentistic] way that other putative [scientific] laws concerning the behavior of such observers under such conditions are generally justified.

-125 "If a relapse into...foundationalism is to be avoided, the [psychological] laws in question must be justified from within the observer's system of beliefs, not by appeal to anything outside it."

126 Justification of premise (2): That conditions C₁ obtain:

-This is justified by a very large number of other beliefs (some of which will themselves be observational and some of which will have the status of background beliefs).

127 Justification of premise (1): That I have a cognitively spontaneous belief P of kind K that there is a red book on the desk:

-This is justified by appeal to the Doxastic Presumption.

-127-132 BonJour offers a lengthy discussion of the justification of the claim that one has a cognitively spontaneous belief of a specified kind which discusses three "sub-premises:" (a) that one has the belief; (b) that it is of the specified kind, and (c) that it is cognitively spontaneous:

--128 (a) [that one has the belief] is supported by the Doxastic Presumption unproblematically--the attempt to question the justifiedness of this sub-premise presupposes the very thing it attempts to question ["How do I know that my belief that I hear a noise is a belief that I hear a noise?" is without sense.]

--(b) [that it is of the specified kind] is also strongly supported by the Doxastic Presumption.

--129-130 (c) [that it is cognitively spontaneous] may be justified by introspection, but (as he will argue below) introspection itself relies upon cognitively spontaneous beliefs.  [130] The justification of this sub-premise is best justified by an appeal to the Doxastic Presumption and [i] "the absence from...[my overall] system [of beliefs] of any beliefs which could serve as plausible premises or intermediate steps for a discursive derivation of the [cognitively spontaneous] belief in question together with [ii] the absence of any positive belief that the [cognitively spontaneous] belief in question was discursively arrived at."

--131 BonJour also maintains that given the "...extreme specificity of content [of the particular beliefs]...it [is] difficult to find premises in one's cognitive system for a plausible discursive derivation of such a belief...and also which marks a belief as being of a kind that is usually or always cognitively spontaneous...."  Contrasting an inferential belief that one will see a milky-white precipitate and the cognitively spontaneous belief that one sees a milky-white precipitate, he notes that: "...what is predicted is a milky-white precipitate in a certain test tube. But what is actually observed is a precipitate which is a certain shade of milky-white, settling in certain specific patterns and swirls, and so on."

6.5 Introspection:

132 Clearly, introspection plays a fundamental role in his justification of observational knowledge.  It is needed to ground claims that a belief is cognitive spontaneous, claims regarding conditions C, and claims that a particular belief token is of kind K.  How are introspective claims justified?

-According to BonJour, the justification of introspective beliefs is treated similarly to the justification of cognitively spontaneous beliefs, although there are fewer premises requisite--the premise regarding "conditions" is, he contends, largely unnecessary.

137 BonJour sums up his discussion of the five-step justificatory argument for cognitively spontaneous observational beliefs by noting that the "...coherentist justification depends essentially on one's reflective grasp of one's system of beliefs; and we have now seen that a coherentist account of observation [also] depends in more specific ways upon that grasp and on the presumption that it is correct.  As I have noted several times, the Doxastic Presumption is only to the effect that one's reflective grasp of one's system of beliefs is approximately correct.  It can now be added that particular questions and issues about the content of one's belief system can be dealt with by appeal to introspection, understood along lines sketched in the present section.  Thus if the issue is whether I do indeed have a particular belief B that I believe myself to have, it is possible to reflect on this question and either have or fail to have an approximate cognitively spontaneous belief.  But of course the justification of this further cognitively spontaneous belief, or of the negative observational knowledge which results from failing to have such a belief, will still depend on my grasp of the rest of my system of beliefs, and thus on the Doxastic Presumption."

138 General summary the coherentistic conception of cognitively spontaneous observational beliefs.

Notes: (click on note number to return to text for the note)

1 Cf., Wilfrid Sellars, "Empiricism and the Philosophy of  Mind," op. cit. 

2 The "type-token" distinction is indicated here.  "Tokens" are specific instances, whereas "types" are general kinds (of which the  tokens are instances).  For example, this annotation is a token of the general type "annotations," and Michael Jordan is a token of the type "basketball players."  

Go to Hauptli's Supplement to Chapter 7

Return to PHI 3300 Home page

File revised on 10/28/2013.