Can a truth value have causal power? A discussion of John Gibbons' "Truth in Action" *
On the basis of some examples John Gibbons tries to show that the notion of similarities and differences between different cases of events reveals the relevance of relational properties, which are of causal relevance. Based on such considerations Gibbons' main claim is that the truth value somebody assigns to his beliefs has causal power. And so the deflationary theory of truth becomes false. So, the questions are: what are (i) similarities and differences between different cases, (ii) what kind of properties are relational properties, and (iii) what is the causal relevance of such relational properties, and why should the truth value be of causal relevance? (iv) Why Gibbons cannot show that the truth value has the relevant causal power?
(i) What are similarities and differences between different cases?
Gibbons assumes events (1) to speak about cases of particular actions and of particular physical states of affairs, however, he does not distinguish them, he simply speaks in both cases about behavior and their causes or about behavioral events.
The example of behavioral events E1 and E2 Gibbons uses is the following:
In short, the similarities of events E1 and E2 are A, aim, B; and the differences are AA, af, O and nb. In other words there are two different owners O of the same B, A and aim; the owners 01 and 02 differ in being different persons with a different near background nb, AA 'how to achieve the same aim', and some irrelevant additional features af.
Based on these two examples of somehow similar events Gibbons formulates his main problem: "The two events are similar in important respects and different in important respects. The relevant questions are in what important respects are the two events similar and different, and, more importantly, what is it for a respect to be important?"(p. 1 )
As I read Gibbons text he suggests that even all the elements that look differently would be adjusted including the assumed sameness of the set of dispositional beliefs, there would remain always one important feature that could never be similar, and this is the set of beliefs that is causally implicated in the action. To get this idea clearer it would be important for Gibbons to distinguish between the causal power of a set of belief and a set of belief that has explanatory causal power. In this version of Gibbons' paper, he seems to shift implicitly between these two notions of causal relevance. However it matters, because of the simple fact that he would have to show that both notions work also in the case, in which two persons have the same mental state, the same set of causally relevant beliefs to achieve the same action, and where Marcia* does and Greg* does not act vice versa. (5)
The answer to the question about similarities and differences of two different events is that in the case that two events including the mental states of the acting persons would be exactly identical (they are only not identical in number) or similar, there would remain at least one feature that cannot be identical, namely the causally relevant set of beliefs. I call this the thesis diff.
Furthermore, with regard to differences Gibbons sketches another version of differences in behavior that is not represented in differences in mental states. In his example af1 and af2 are differences in the behavior of Marcia and Greg, where "Marcia casts a shadow in my direction, while Greg does not. But like Greg, she has no thoughts at all about me, my direction, or her shadow."(p.2) Gibbons explains the difference in behavior on the claim that Marcia's casting the shadow is not something she did intentionally, such that 'casting a shadow...' cannot be mentally represented. Two objections can be formulated: how does Gibbons know that it is not performed intentionally, because it could be the case that she met Gibbons some time ago and now she recognized him, and so she intentionally casts a shadow toward Gibbons. Gibbons himself has a mental representation of the casting a shadow, but as he believes that it is not performed intentionally, he behaves according to his belief that it is not performed intentionally, however, he himself might have behaved differently if he had taken this casting a shadow intentionally. For the reason that he does not know whether or not this casting a shadow toward him is performed intentionally or not, he cannot explain the difference on the assumption of intentionality and its mental representation. On the other hand, Greg could intentionally not cast a shadow toward Gibbons, because of something we do not know, and so Greg has according to Gibbons characterization of doing something intentionally a mental representation of not casting a shadow toward Gibbons. (6) The second objection is that from the fact that Marcia actually might not have an explicit thought about her casting a shadow she might have implicit thoughts. For example during she has to drive her car, to stop it because of certain reasons, her mind might not have been occupied by such stupid thoughts, she might think something else about the concert she attends this evening: tatatata...tatatata... the break in Schostakovitschs 1 st violincello concerto is beautiful ..ta..but it is otherwise tatttta...a..a...a...tttttt... Or such things of this kind. The point is that she has a mental representation of something that looks like casting a shadow toward Gibbons. However, Gibbons cannot find out what it is, because he has no direct look into Marcia's mind and thereby to the mental representation, but if he could have such a look, he would find according to his characterization of mental representation (7) the mental representation of tatatata...tatatata... the break in Schostakovitschs 1 st violincello concerto is beautiful ..ta..but it is otherwise tatttta...a..a...a...tttttt... Would Marcia's correlated behavior that looks like she would cast a shadow toward Gibbons be interpreted differently? The important is that Gibbons could not say that the behavior does not have a mental representation.
The problem Gibbons cannot solve here is the notion of the relationship between mental representation and doing something intentionally, therefore, I note that this relationship remains entirely unexplained. A further problem is that he seems to assume that persons can have only one thought at once in the mind, step by step, but this assumption is obviously false. The notion of the individuation of behavioral events on intentional ground also does not help him to save the exclusive relationship between doing something intentionally and mental representation, "involving the relevant action type."(p.2) This is simply begging the question. It is interesting that at this point of the discussion Gibbons moves to the weaker notion of causation, to the explanatory causal role of sets of beliefs. I quote:
What Gibbons wants is an explanation of an action, which includes the notion of a systematic or lawlike connection between mental states and behavioral events. I call this the thesis sys. For this purpose he assumes a particular description such that the action appears to be intentional under that description, and only the actions described in such a way can have a mental representation. To find such systematic or lawlike connections seems to be Gibbons' scientific interest. However, the problem is that his characterization of doing something intentionally cannot be helpful for such a purpose. Neither can the notion of the individuation help him: mental states are either causally related to behavioral events in the relevant way, and so there exists a lawlike connection, or there does not exist such a lawlike connection because of the fact that what counts as behavior is all that is explicit of implicit intended by the person who behaves somehow. Therefore, the notion of doing something intentionally seems to contradict the notion of such a lawlike or systematic relationship between states of the mind and behavior or particular behavioral events. Furthermore, there is no need that mental states are individuated intentionally, and so the relationship between mental states and intentions itself remains unexplained. So, "behavioral events" cannot be individuated intentionally. On Gibbons' account it would be sufficient to assume that "behavioral events" are individuated by the notion of an event. From such a notion of an individual "behavioral event" Gibbons could construct a lawlike relationship to mental states. The particular mental event that causes the particular behavioral event (8) would be the causally relevant element. Instead of trying to solve this problem, Gibbons moves to something else, and adds his arguments for the thesis diff, according to which at least there remains a difference in the set of beliefs that are causally relevant, by which this set of beliefs is individuated by intentions, and for thesis sys , namely to considerations about the use of natural languages and everyday explanatory practice of actions. I quote:
By these considerations about natural languages Gibbons falls simply the epistemological aspect. Doing something somehow perceptual beliefs or perceptual knowledge (9) and knowing how to do something can rather be described as an automated process than a typical case for an explicit formulated set of beliefs hold to be true by a particular person at a particular time and at a particular place. To describe this issue as a mistake of the use of our natural language cannot count as justification, and so the notion that there should be something below the surface of our ordinary explanatory practice. (Compare p.3) Gibbons mentions that ordinary explanations are simply incomplete with regard to the reconstruction of the relevant causal connection between a state of the mind and behavior. And even the ordinary practice of explanation would be mistaken, Gibbons could not derive or clarify from such a fact an appropriate solution of the relationship between intentionally, mental states and behavior. Consider the following sentence: Marcia intended to stop at the red light, but she did not stop. There is no systematic relationship between the intention to act, to intend something, and an act, because all that might be intended could not be performed, or it could be intended not to act. So, there are mental states, but there is no corresponding act. Or there is a mental representation of a non-act or of omitting an act, etc. Gibbons cannot save his thesis diff and thesis sys on his considerations about a wrong practice of the use of our natural language and everyday explanatory practice.
(ii) What kind of properties are relational properties?
It is not easy to understand Gibbons move from the causal relation between events and the causal relation between properties. The motivation becomes not clear, although Gibbons mentions that otherwise it would not be easy to understand Mill's methods of difference and identity. For this purpose Gibbons says that for an application of Mill's methods we cannot work with nonrepeatable features. Events are not repeatable. Therefore, Gibbons is looking for repeatable features that could reveal the causal relevance within some test. To describe the problem Gibbons gives an example of a physical event of a play with a pool in a bar. The relevant feature is that the ball comes back if it is white, and it will not come back it is not white. What is the relevant feature for this regularity? Gibbons shifts to an anthromorphism when he asks "but we might also want to know how the table knows that it's the white ball and not the black ball. Or, to put the question less metaphorically and more metaphysically, what are the causally relevant features of the cue ball that account for the difference in its behavior?"(p.5)
I suggest that Gibbons should rather speak of regularities here instead of the behavior of the ball. Physical regularities have some causes that can be described in terms of physics, and so the notion of a metaphysical aspect is misleading. Both is misleading in the description of this physical regularity, the metaphorical and the metaphysical question. Which one of all the properties of the ball is relevant to guarantee the regularity in question? Is it the color, the weight, the mass, the size, or something else? There is no metaphysical question to answer, but simply a physical one, namely: what is the best and/or cheapest method to construct such a physical regularity? Different methods are available, and which one we choose is a matter of our decision. Then, the causally relevant feature of that physical regularity depends on the physical method we have chosen before, and which we have used for the construction of such a regularity. A counterfactual analysis of the following kind is not helpful: "if the ball that went in had been white, it would have come back out."(p.7) There is an element missing: it is the physical construction that guarantees such and such a regularity. So, the causally relevant feature of this regularity can be described relative to the method of the construction of such a regularity. Besides that Gibbons could not show that there do not remain non-relative genuine features, namely the laws of physics that are exploited to construct such a regularity. The description of such a regularity cannot reveal the causally relevant facts or features of that regularity. I claim that the description, although in counterfactual conditions, is of no importance with respect to the causally relevant facts of that regularity. Based on the considerations above there is no need to discuss Gibbons arguments against the deflationary conception of truth, because Gibbons arguments are insufficient to support his thesis. Gibbons provides us simply with unjustified ad hoc arguments. It is rather the case that Gibbons has to show that these physical regularities need a metaphysical or metaphorical interpretation he stated. It is still Gibbons burden to show that there is something metaphysical within physical regularities. How to describe such physical regularities with respect to the causally relevant features also does not involve a metaphysical description. I think that Gibbons way to analyze physical regularities is mistaken.
Nevertheless, based on the consideration of metaphysical aspects within physical regularities, Gibbons comes back to the cases of human actions. He takes the physical properties of the ball to be in some sense similar to the intentional properties of behavioral events. (10)
The answer to (ii) is that the relevant properties are relational properties with respect to a particular method and an effect intended by somebody, and these are the relational and intentional properties in question. With respect to human actions Gibbons assumes intentional behavioral properties that are the relevant relational properties relative to a certain effect intended. However, Gibbons did not derive such a kind of properties from an investigation of human behavior or epistemological considerations, but from a metaphysical interpretation of a physical regularity.
(iii) What is the causal relevance of such relational properties and why is the truth value such a property?
What counts to be causally relevant can be described relative to a physical method used for getting the physical regularity or effect in question. Based on such a notion of a relative aspect with regard to a particular method and a certain effect, Gibbons claims: "this looks like a case where a relational property rather than some intrinsic correlate of that property is "doing the causal work"."(p.6) The line of reasoning here is far from being justified and clear. However, from the notion of physical relational properties, instead of intrinsic properties, Gibbons simply shifts to a certain notion of intentional behavioral properties, which are important for the assumption of thesis sys based on thesis diff, and then he states the following:
"Now I claim that the truth value, not just truth conditions or content of your beliefs is causally relevant. Truth is a causal power."(p.6)
This is really a strong claim. Nevertheless, what is Gibbons' justification? Gibbons applies the property of being true to beliefs. But it is far from clear how the property of being true is related to the assumed intentional properties of behavior. The counterfactual fact Gibbons proposed cannot show that a truth value is connected to intentional properties of the relevant kind. With respect to our example: Gibbons changed something, namely the example of event E in which the belief B1 is that Marcia thought that stopping would help to ovoid getting a ticket into the example of event E*, in which the aim1* is 'wants to stop at the red light', and where the belief B1* 'is believes that by squeezing this lever, she will stop.' From E* Gibbons concludes "if her belief is true, she will stop. If the belief has been false, she wouldn't have stopped. This suggests that truth is causally relevant to ordinary end result types like stopping at a red light as well as their intentional counterparts."(p.6) So, B1* is not part of event E, because in E Marcia and Greg intended to ovoid getting a ticket according to aim1 and aim2. Relative to the intended aim1 and aim2 or the intended effect there is a belief B1 and belief B2 that stopping would help to ovoid one. These beliefs B1 and B2 are individuated by their owners 01 and 02. However, B1* says something different, namely that the intended aim* was stopping at the red light. This is not true in E, and so B1* cannot be part of E, but is part of E*. In E* the aim* or effect of the action A* is the pure action A1* or A2* itself. The consequence is that in E* the action A* and the intended aim* are identical. The belief B* is reduced to the additional action AA* to achieve action A* that is identical with aim*. It is easily to see that Gibbons simply says that each action is selfjustified, because otherwise it would not have occurred. In the case that there is some effect, there must be a lawlike or systematic correlation to a set of belief represented in the brain hold to be true by a particular person. However, from Gibbons selfjustified example E* he cannot conclude that there exists something that could be called a relational intentional behavioral property relative to the effect produced, which is the causally relevant truth-value.
(iv) Why Gibbons cannot show that the truth value has the relevant causal power?
Now, my question is: can Gibbons at least show that the truth value is somehow causally relevant in E*? I tried to figure out that the logical structure of E* is a selfjustifying structure the intended effect or aim* of an action A* is identical with aim*, and so the belief B* is reduced to the additional action AA* that is to know how to do something, and as already pointed out B* is simply a perceptual belief neither true nor false. How does the notion of the truth value enter the example E*? The truth value is applied to the perceptual belief B* that by squeezing this lever, she will stop. However, as already mentioned the notion of knowing how to do something does not involve the notion of explicit beliefs held to be true by a particular person at a particular time. Therefore, my objection against the claim that the truth value itself is causally relevant in an action of the type E* is not justified and false. The crucial point is that contrary to Gibbons' claims there does not exist such a thing like a causally relevant property defined relative to an aim of an effect that is the truth value. In E* the notion of a causally relevant truth value simply does not exist. Besides this fact, sometimes something non-existing might turn out not to be relevant with regard to other aspects not included in the notion of perceptual beliefs.
I conclude, the reason for my claim is that in Gibbons' example of the event E* the intended aim* is identical with the action A* itself. In E* in the case that the person acts, she simply acts, and so the notion of the truth value cannot have causal power, because in the negative case, where she did not act, we could not say that there was the truth value 'false' assigned to the same belief, and so she did not act. There cannot be a truth value at all that might be causally relevant with respect to a perceptual belief B*. To know how to do something is not a matter of explicit beliefs held to be true or false by a person when she acts. (11)
In short, my objection against John Gibbons's considerations is that he simply cannot show that the truth value determines causally human actions in the relevant sense. However, I think that there might be other ways to include a somehow notion of truth into everyday human behavior not considered in John Gibbons' paper on "Truth in Action".
Davidson, Donald: Mental Events; in: Essays on Actions and Events, Oxford: Clarendon Press 1980.
Gibbons, John: Truth in Action.
Lehrer, Keith: Theory of Knowledge, Boulder, San francisco: Westview Press 1990.
Papst, Josefine: Vernuenftige Wesen, ihr Wille und ihre Freiheit. (unpublished)
Pollock, John: Contemproary Theories of Knowledge, Maryland: Rowman & Littlefied 1986.
* For discussions and comments on this paper I am grateful to John Gibbons and Thomas Nagel.
(1) The notion of events is based on Donald Davidson's analysis of causation. Causation is event causation. Compare Donald Davidson: Mental Events; in: Essays on Actions and Events. Oxford: Clarendon Press 1980. With the notion of events Donald Davidson rejected Quine's view according to which ontology remains dark, such that we can decide whatever we want without being able to justify our decision. Other candidates that could be taken to be the elementary entities on ontological grounds are properties, objects, facts, situations. However, whatever we assume to be the most elementary entity, we will get problems of different kinds and relevance. Yet in this paper the problem of the most considerable elementary entities is not discussed. Gibbons simply uses the notion of events to speak about particular cases.
(2) Davidson discusses such examples, and asks whether or not there are basic actions that are most basic. In Gibbons example, the question is: was the basic action stopping or squeezing a lever or pressing a pedal, or something else, such as some mental or physical state, and the causally relevant belief about it, to which in the end a truth value is assigned? I come back to this when the causal power of the "truth value" is discussed.
(3) There might be a bunch of additional features that could be mentioned, such as those Gibbons describes, for instance which car they drive, how they move, etc., and we can continue the list with: which sentence was spoken to somebody, what were the lines in the faces, how somebody looked like, what the mood of somebody was, etc. There could be given a long list of additional features of an action. Compare for instance the first sentence of James Joyes "Ulysses". A thought is expressed by one sentence, so, in the case that somebody has a complex thought, the sentence might become long, and so the assignment of a truth value might be difficult or impossible. As in Joyes' case 26 pages. The question is: where to put a point to finish a sentence, although the thought would still not be finished. The point would end the unfinished thought expressed by a finished sentence, but could the point really finish the unfinished thought? Where to put a point? It seems to be a matter of preference. In other words: where should we put the end of the list of additional features of an action A? And how do we know that some of the additional features are not relevant for the evaluation of the beliefs about the intentions regarded to be true.
(4) There is no reason to assume only one belief to get the aim, because there might be alternatives to avoid getting a ticket. Therefore, there must also be a belief about the alternatives and so forth.
(5) Here, I introduce Marcia* and Greg*. The difference-between Marcia and Greg, and Marcia* and Greg* is that in the latter case they both are perfect twins, including their brain states and the causally relevant set of belief. They differ only in one feature: where one* acts the other* does not. I think that this is a serious counter example for Gibbons. Applied to Gibbons assumptions it would follow: (i) in the case that he assumes a set of belief to be causally powerful, he would have to reject the idea that there is a set of such causally relevant beliefs, or (ii) he can use the notion or explanatory causally relevant, but without causal power at all.
(6) So, Gibbons assumption that Greg's behavior of not c-casting a shadow toward Gibbons, and his possible behavior of not intentionally casting a shadow toward Gibbons, and his possible behavior of intentionally not casting a shadow towards Gibbons cannot have a mental representation of the relevant type is false.
(7) The notion of mental representation is tricky, nonetheless, this issue should not be discussed here.
(8) It is important to mention here that Gibbons account of behavioral events and their causal relationship to mental events or state is quite similar to the account Donald Davidson gives. According to his anomal monism Davidson assumes that there are mental token events that correspond causally to a particular physical token that is behavior. Davidson was not able to show that there really exist such a causal relationship between mental events and physical events (behavior). Compare also my paper on this topic; Josefine Papst: Vernuenftige Wesen, ihr Wille und ihre Freiheit. (Unpublished)
(9) Such problems are well discussed for example by Keith Lehrer and John Pollock. Compare Keith Lehrer: Theory of Knowledge. Boulder, San Francisco: Westview Press 1990. Especially chap. 4: Fallible Foundations; section on Perceptual Belief and Independent Information. Keith Lehrer: Der Quadrant der Vernunft. (Unpublished). Keith Lehrer: Knowledge, Coherence and Skepticism. (Unpublished) John Pollock: Contemporary Theories of Knowledge. Maryland: Rowman & Littlefield 1986. Especially chap. 2; section 3.2 The Problem of Perception; and chap. 3, section 5. Nondoxastic Theories and Direct Realism. John Pollock: Epistemic Norms; in: Synthese 71, 1987, pp.61-95.
(10) It is still not clear for what reason Gibbons always considers Mill's methods to justify his claim about the metaphysics of causally relevant properties.
(11) Compare Pollock's discussion of the intellectualist modal. Pollock believes that it is wrong or only relevant in the negative case. Compare again John Pollock: Epistemic Norms, in: Synthese 71, 1987, pp.61-95.