Susanna Siegel

Susanna Siegel (Harvard University), “The Visual Experience of Causation,” with commentary by Sarah McGrath (Brandeis University).  The paper can be found here.

Comments

[Christian Lee]

Dear Susanna,

Very interesting read. I have been thinking about this issue a bit lately and although I can't justly claim to know the literature well, I have some worries and questions. Here goes...

1. It seems to me very important to be clear as to whether one is defending a direct or indirect realist view of sensory experience. A variety of points you raise in the paper seem to me only to be consistent with a direct realist viewpoint. I have in mind the idea that one is sometimes directly aware of external objects in perception. This, if true, does diminish the strength of the argument.

2. One could take experience to be a relation between an individual and an object at a time. If this is correct, then it is not obvious to me how one can experience causation between events at different times. As an example, suppose Presentism is correct and that the present is instantaneous (or rather short). If so and seeing is factive, then one will run into trouble making sense of the view that one sees a causal relation between non-cotemporaneous events. Of course, causation is problematic for a Presentist, but it seems even more so for the view you are arguing for.

3. Your strategy is an inference to the best explanation in that you think experiencing causation best explains the phenomology of an experience that involves causation contrasted with one that does not. I have a worry about this strategy. First, it isn't clear to me that when I assert that "it seems to me that P" in the relevant cases, what I am reporting is the content of my experience. I could just as well be reporting something inferred from my experience. No doubt we wouldn't know the difference. Second, if indirect realism is true, then this going to be an attractive account of just what one is doing when one reports causation.

4.You seem to assume that experience represents motion, but again, this will not be implied by an indirect realist view. I am also skeptical about this claim working as an assumption in your argument since it seems to me as, or nearly as, debatable as the view that causation is represented in experience.

5. You also seem to think that temporal relations, like simultaneity, can be represented in experience. This is also very debatable, especially if one thinks it necessary that presentness is instantiated by a possible experience and that it cannot be co-instantiated (intrinsically) with other tensed properties. This should make us weary of bringing in 'earlier than' and 'later than' relations into the content of an experience.

6. The idea of unity comes up quite a bit in that it seems you think some experiences have a peculiar kind of unity that other experiences lack. This seems to function as intuitive data you want to confirm by appealing to causation. Two things: On an indirect realist view this seems to undermine the argument. One could have indistinguishable experiences, one involving causation, and the other not, and it said that the content of the experience is what the experiences have in common, that with which one could be directly aware in both experiences. But direct awareness is factive and causation is missing in one case so it will not be part of the content. One might argue that there is extended content, properties and relations with which one is not directly aware, but then it will need to be argued that this added content is not inferred or content of a different state. That leads up to my second worry: perhaps experience of causation is a complex state incorporating a memory and an sensory experience causally related to one another in a certain way. Roughly, one sees the curtain at a certain location, remembers it being in a slightly different location, then sees it in a near location, then remembers being....so that memory and experience jointly cause the phenomenolgy of causation. Perhaps attitudes of expectation and background beliefs will be part of the story as well. This strikes me as an attractive view that should be considered.

7. Similarly, perhaps the phenomenology of causation has its immediacy due in large measure to the immediacy of the memory. The memory will be very immediate in this case, of course, but I am very doubtful that one would be able to introspectively decide whether experience of motion or causation did or did not involve memory.

8. In the ball turning off the light example you appeal to the idea that one experience is falsidical and the other not, on the assumption that the ball does turn off the light in one case and not the other. I'm inclined to think this data can be explained just as well be referring to the belief one forms being true in one case and false in the other.

9. I like the strategy of arguing that appealing to a disposition to form a causal belief gets the order of explanation backwards, that is, it is the content of the experience that explains why a disposition to believe a causal relation obtains, is formed rather than not. Even if this were right, I don't think this would tell against the account I sketched above.

10. You appeal to cases of "what is seen" to respond to worries and develop your view. But I think I can see the president of Zambia and not see him as the president. Seeing does not seem to me a plausible relation to motivate views about content, here direct awareness does better I think.

11. You suggest that one can represent causation without representing a distinctive feature of it. I wonder if this is consistent with the kinds of cases you describe to motivate your view? For example, can I experience causation thus representing it without there seeming to me to be unity in my experience?

12.Lastly, the necessary condition you end with seems to me somewhat weak and I wonder what work it might do. To represent two things (at a time) one must represent them as contiguous (spatio-temporally related). Could one experience any two objects or properties at a time and fail this condition?

Okay, that was longish I know. Like I said, I've been think about this stuff lately and was eager to get your take on some things.

[susanna]

hi christian, thanks very much for the twelve questions. here are some replies to 1, 2, 3, 4, 5, 8, 11, and 12. i’ll save replies to 6, 7, 9, and 10 for tomorrow.

re. 1: At the start of my paper I distinguish the question whether causation can be represented in experience, from the question whether causation can be perceived. I mostly focused on the question about experience. i’m not sure exactly what you have in mind by ‘direct realism’ and ‘indirect realism’, since there are a variety of kinds of directness and indirectness. if we work with a theory on which perception is indirect if you perceive external objects by perceiving non-external objects, then the theory that experiences have accuracy conditions is neutral on whether perception is indirect. So I don’t see that my central argument for the main thesis takes a stand on that issue.

re. 2: I don’t take experience to be a relation between an individual (I take it you mean, the subject of the experience) and an object at a time. If there is a problem about combining Presentism with the claim that causation can be perceived, it does not seem to affect whether causation is represented in experience.

re 3: I agree with you that we can’t take ordinary uses of ‘it seems to me that p’ as a guide to the contents of experience. I make this point in the section on michotte. I can’t endorse your description of my strategy though, since the way you put it, the explanandum is ‘the phenomenology of an experience that involves causation’. Clearly ‘involves causation’ here could not reasonably the same as ‘represents causation’, since if it did there wouldn’t be any need for the putative explanation, since what the strategy is employed to show is exactly that experience represents causation. So in your gloss you seem to be taking ‘involves causation’ to mean the same as something like, ‘is reported using causal terms’. that is not how I introduced the explanandum. I introduced by describing a familiar kind of experience. It’s true that we might in fact use causal terms to describe this kind of experience, but I wasn’t resting much weight on that. i certainly wasn’t assuming that that showed anything about the contents of the experience reported – if I did, i wouldn’t have to bother comparing the various possible explanations.

re. 4: You’re right that I was assuming that experiences can represent motion. Why do you think indirect realism has to deny this? I don’t see why you think it’s as controversial as the main thesis. experience might represent that object A moves and that subsequently object B moves. even if their movements are in fact a case of causation, it won’t follow that the experience represents causation. did you have a specific debate in mind about whether motion is represented in experience?

re. 5: You’re right that I was assuming that temporal relations can be represented in experience. What’s represented in experience is constrained by the phenomenal character of experience, and it seems plain to me that just as something can look red, so too it can look like it’s moving at the same time (or just before or just after) something else. If Presentism is true, then some experiences may turn out to be inaccurate. But it won’t be a bar to having such an experience in the first place.

If the only sort of experience you thought could be had were experiences that consisted in a (perceptual) relation to an object and a time, then you wouldn’t allow this move. Maybe that’s what you had in mind when you said I was assuming indirect realism: maybe you meant that I was assuming that falsity of direct realism, where direct realism is a view about the nature of experience like the one Mike Martin holds (a variety of disjunctivism). If so, I think it would be useful to distinguish two versions of direct realism. There is the version just mentioned, championed by some disjunctivists, which is a theory of the nature of phenomenal experience itself. Then there the negative claim about perception (nb: not necessarily about phenomenal experience itself), that when we perceive external objects we don’t do so by perceiving internal or mental objects. If you’re a direct realist of the first (Mike Martin-esque) sort, you’ll also be a direct realist of the other sort. but the converse does not hold.

Everything I say in my paper is neutral on whether direct realism about perception is true. Some things i say in my paper rule out Mike Martin-esque direct realism, where this is the claim that the only sort of experience that can be had are experiences that consist in a perceptual relation to an object and a time. (that’s not Martin’s own formulation but it is close enough to be martine-esque). This kind of direct realism says that the only kind of metaphysical structure had by experiences is a structure of perception. I think that’s false – some experiences are structured like propositional attitudes. Of course anyone who holds what I've taken the liberty of calling a Martin-esque position (please forgive me, Mike!) has to have something to say about what is happening in cases of hallucination. It is the troubles they have with this task that lead me away from that view.

Even though I assume in my paper that visual experiences are structured like propositional attitudes, something very much like the central question of my paper- whether causation is represented in experience – can be posed even if the assumption is false. For instance one can ask, which properties or relations is one presented with, when one stands in the Martin-esque relation to an object that constitutes your having an experience? The main strategy of seeking the best explanation for phenomenal contrasts can still be applied. I suspect that there will be analogs of the arguments within the framework that do away with accuracy conditions for experience, though I haven't thought completely through had that would go for each alternative to that framework.

re. 8: I don’t think the data can be explained just as well by referring to the belief, since there will be cases where you don’t form any such belief. For example, say you flick a switch in your apartment and the lights in the Eiffel tower go on. You might not believe that you turned on the lights, but it might nonetheless in some sense seem to you that your flicking the switch had turned them on.

re 11 (“You suggest that one can represent causation without representing a distinctive feature of it. I wonder if this is consistent with the kinds of cases you describe to motivate your view? For example, can I experience causation thus representing it without there seeming to me to be unity in my experience?”): I think it is consistent, even if we assume that some kind of unity is a distinctive feature of causation. The point of the holes section is not that, for every distinctive feature D of property X, you can represent X in experience even if you don’t represent D. The point was rather that from the fact that D is a distinctive feature of X, it does not follow that if you represent X, you represent D. I am responding to an opponent who says, “we don’t represent counterfactual dependence [or some other putative distinctive feature] in experience; but counterfactual dependence [or the instantiation of general laws...] is a distinctive feature of causation; therefore we don’t represent causation in experience.”

re. 12 (on apparent spatio-temporal contiguity as a phenomenal constraint on experiential representations of causation) : it was only supposed to be a necessary condition.

[Christian Lee]

Susanna, let me fill out in a bit more detail a case to motivate my comments 2, 3, 4 and 5.

Suppose one sees before them a ball and it is red all over at t1. Immediately thereafter, the ball changes to being green all over at t2. Question: does one thereby have an experience that represents a ball with incompatible properties? I think the answer is no. There are two ways to explain this. First, I have an experience that represents that the ball is red all over at t1 and green all over at t2. I represent an object that instantiates compatible properties. Second, I represent at t1 that the ball is red at t1 and I represent at t2 that the ball is green all over at t2. I have two different experiences, and there is no single experience which represents the ball as being incompatibly colored at a time.

There are reasons to favor the second explanation. For one, the first explanation entails that one represents conjunction in experience. But one does not represent conjunction in experience. So, the first explanation will not work. Two kinds of ideas supporting the claim that conjunction is not represented in experience are that (i) if conjunction, then why not disjunction and negation, but not disjunction and negation, and (ii) conjunction is not the kind of property one can be sensorily aware of so it is not the kind of property that can be represented in visual experience. So, I prefer explanation 2.

If explanation 2 is correct, then we need to consider experiences as of a particular time and then ask: what is represented in experience at that time. Suppose that the time needs to be instantaneous. It follows that the contents of that experience will be instantaneous. But, then the worry is that this excludes causation from being represented in a particular experience because causation "takes time". More generally, if C causes E, then E is later than C. If E is later than C, then E and C cannot both be represented by a sinlge experience. Similarly, the reason why being red all over at t1 and being green all over t2 is not represented in experience is that there is no single experience that represents both of these properties.

This point is general and it will also exclude motion from being represented at a time because motion "takes time". We could introduce a term, representation* that is conjunction of representations to capture the idea that motion is represented in experience. I may represent at t1 an object at a location L1 at t and represent at t2 the same object at a location L2 at t2, and so on. But representation* is not representation. Anyway, if this right, then there will need to be a non-trivial restriction in the thesis that causation is represented in experience. The relevant causation will need to be simultaneous causation and it isn't clear whether this weaker thesis is sufficient to establish that causation is represented in experience or sufficient to establish the modes of causation you discuss.

Onto the "Indirect Realism". I suggest that your paper seems to assume that Direct Realism is true. There are a number of reasons for this. For example, you say that during visual experience one represents that something is case. Would an IRist say this? Maybe not. She might think that what is represented in experience are privately accessible properties and their intrinsic relations. She might argue that they do not stand in causal relations to one another, but rather, in between-ness relations, for example. She will go on to argue that it must be inferred that there are external objects that cause the relevant private properties to exist and stand in the relations they appear to stand in and also that there are external objects that stand in causal relations to each other. But, since inference is involved, what is inferred in not part of the content of the experience. This view may also (likely will) imply that we do not represent tables and chairs in experience, but that we infer them as the causes of one's experiential content.

This is all supposed to be consistent with the fact that when someone reports her experience she correctly reports it saying, "It seems to me that there is a table in front of me." The idea is that what is reported is a belief (assertive attitude) that is inferred from the content of the experience.

This all stands in stark contrast to your initial claim that representation is not factive. On this view, representation is factive. The representational content of an experience is exhausted by the properties, objects and relations with which one is directly aware or directly acquainted. And direct awareness is factive. IRist will argue that the content of the experience is what veridical and falsidical experience would have in common. Or, in the form of an argument, one might say that if IR is true, then the content of an experience is that with which one is directly aware and since we represent tables in experience, then the representational content of an experience is not part of the "sensory content" of an experience, instead, it is the content of what is inferred from the sensory content of an experience.

If this is what Irist say, then you are assuming that IR is false, right?

[susanna]

Hi Christian, I still owe you answers to a few of the original 12. Meanwhile here are some replies to the first half of what you say above.

“Suppose one sees before them a ball and it is red all over at t1. Immediately thereafter, the ball changes to being green all over at t2. Question: does one thereby have an experience that represents a ball with incompatible properties? I think the answer is no.”

I think the case hasn’t been described fully enough for there to be answer. Some worldly stimuli have been described, but nothing has been said about how those stimuli appear to the subject.

“There are two ways to explain this. First, I have an experience that represents that the ball is red all over at t1 and green all over at t2. I represent an object that instantiates compatible properties. Second, I represent at t1 that the ball is red at t1 and I represent at t2 that the ball is green all over at t2. I have two different experiences, and there is no single experience which represents the ball as being incompatibly colored at a time.
There are reasons to favor the second explanation. For one, the first explanation entails that one represents conjunction in experience. But one does not represent conjunction in experience.”

I’m not sure whether we disagree about representing conjunction or mean different things by it. Let’s take conjunction between predicates. All I mean by the claim that experience can represent conjunction between predicates is that for some properties F and G, an experience can tell you that an object has F and G. If the object doesn’t have F or G, then what your expeirence tells you is false. For instance, I might see a white cube, call it C, in virtue of having a visual experience as of a red cube. Such an experience would be falsidical. It’s falsidical, because it’s telling me that C is red and a cube when really it is white and a cube. Maybe you’d deny that it’s falsidical. I think that’d be odd, since it seems like a straighforward case of illusion. If the experience is falsidical for the reason described, then it is taking a stand on whether C has the properties of being red and cubical. So it is ‘representing conjunction’. The same case seems to give us two propositions. I think the experience described represents that C is a cube, and that C is red. The second proposition is false, so the experience is falsidical.

“So, the first explanation will not work. Two kinds of ideas supporting the claim that conjunction is not represented in experience are that (i) if conjunction, then why not disjunction and negation, but not disjunction and negation, and (ii) conjunction is not the kind of property one can be sensorily aware of so it is not the kind of property that can be represented in visual experience. So, I prefer explanation 2.”

Re (i), the reason the experience in the example doesn’t represent that C are blue or red is that there is no phenomenal character that such a disjunction would reflect that isn’t already reflected by representing that C is red. For the case of negation, I think visual experiences can sometimes represent absences. You might think of that as a kind of negation. I discuss that case in my entry on “The Contents of Perception” in the Stanford Encyclopedia. Roy Sorensen also has a nice discussion of the general topic of perceiving absences in a forthcoming book.

Re (ii), I don’t think that when we decide whether experiences represent properties, we take the properties one by one and try to ascertain whether one could be sensorily aware of them. Maybe that reflects some sort of differnece in what we mean by ‘represent properties’. I think the basic notion is the notion of experiential contents (where these are given by accuracy conditions that reflect the phenomenal character of the experience). Which properties an experience represents depends on which accuracy conditions it has.

re. simultaneous causation: I discuss some cases of it – eg the cat weighing down the hammock. My central thesis will be true if experience represents that the cat is weighing down the hammock (even if it doesn’t represent that the cat is a cat or that the hammock is a hammock).

[susanna]

Hi again Christian, here are replies to what you say about Indirect Realism.

You say,

“I suggest that your paper seems to assume that Direct Realism is true. There are a number of reasons for this. For example, you say that during visual experience one represents that something is case.”

Thanks, this helps me see how you were thinking about things. But you seem to be conflating the claim that experiences have contents given by accuracy conditions with a more specific claim about what those accuracy conditions are. You could be an IR-ist (e.g., someone who thinks we see external objects by seeing mental sense-data) but still hold that experience has accuracy conditions that concern the sense-data rather than external things, or that it has accuracy conditions that concern sense-data in addition to having ones that concern external things. Logical space even has room for a position that says that experience represents causal relations between sense-data. Maybe such experiences would always be inaccurate (if sense-data can’t cause things), but that does not stop them from having accuracy conditions.

"...This all stands in stark contrast to your initial claim that representation is not factive. On this view, representation is factive.”

Here it sounds a bit as if you think there is a single phenomenon, experiential representation, and the IR-ist and I disagree about its nature. But the notion of representing something in experience is a technical notion, not part of the phenomenon that theorists are giving competing accounts of.

What you’re pointing out is that there is a position in logical space that invokes a factive notion of representation and says that (i) experience represents that mental items have certain properties (and relations), and (ii) if experience represents that a mental item has property F, then that mental item has property F.

I’m assuming that experiences represent that objects in the external world have certain properties. But that assumption isn’t playing any crucial role in my argument.

The notion of representation in experience that I’m using is tied to the veridicality and falsidicality of experience. The notion of representation that you’re using, in contrast, seems to come apart from those notions since you say later, “the IRist will argue that the content of the experience is what veridical and falsidical experience would have in common”. Since the notion you’re talking about is factive, an experience won’t count as falsidical in virtue of its wrongly representing that something obtains. That raises the question of what you mean by ‘falsidical’.

[Christian Lee]

Hi,

"I think the case hasn’t been described fully enough for there to be answer. Some worldly stimuli have been described, but nothing has been said about how those stimuli appear to the subject."

Suppose that instead of using "sees" I used "appears" instead. So the case would be one in which: It it appears to one that there is a ball and it is red all over at t1. Immediately thereafter, it appears to one that the there is a ball that is green all over at t2. Also stipulate that the subject has an idenficatory background belief so that she would report the second experience as: it appears to me that the very same ball that was red all over at t1 appears to me to be green all over at t2.

"All I mean by the claim that experience can represent conjunction between predicates is that for some properties F and G, an experience can tell you that an object has F and G."

What worries me here is that experience does "not" tell you that the ball is red and later green even though one would report the experience by saying "It appears to me that the ball is red and then later green." The idea is that one's reports of how things appear can be misleading. On the one hand, they can be taken to report the content of the experience. On the other hand, they can be taken to report some content that "downstream" from the experience itself. This content will be inferred from the content of the experience, but not actually part of the content of the experience.

So, if it appears to me that there is a red cube, then I would say that the content of the experience (what one is directly aware of during the experience) is an instance of whiteness and instance of cubeness and perhaps a visual field as a backdrop. So, this would be the content of the experience and it's existence is not sufficient (I'm suggesting) for one to be directly aware of conjunction even though one would report her experience by saying "It appears to me that there is something that is red "and" cubical." What one reports is content that is "downstream" from the experience itself.

I don't then deny that the experience isn't in some important sense falsidical if a white cube ones sees appears to one to be red. I'm claiming that the experience will be falsidical in this sense: representational content that is inferred from the content of the experience does not correspond to the way the world is. Now, I actually think that something better can be said. One could say the representational content inferred from the experience is falsidical because it has propositional content that is not true and the content of the experience is "falsidical" because it does not map the environment. Here "mapping" is like representing except for the fact that (i) its not propositional and satsified in the way propositions are satisfied and (ii) it is to be cashed out in terms of systematic counterfactual dependence between properties and relations with which one is directly aware and properties and relations that external objects stand in to one another.

I'm not saying mapping is representing though, rather it is a relation that an Indirect Realist will appeal to to account for the intuition (if there is one) that the content of an experience can be more or less accurate.

"The reason the experience in the example doesn’t represent that C is blue or red is that there is no phenomenal character that such a disjunction would reflect that isn’t already reflected by representing that C is red. For the case of negation, I think visual experiences can sometimes represent absences."

I'm not sure I quite understand. Is the idea that in general, two states are representational the same when they have the same phenomenal character? Representing that C is blue has the same phenomenal character as representing that C is blue or red. So, they have the same phenomenal character. By itself this wouldn't show that disjunction is not represented though right? Perhaps you are happy with a reductive assumption here. If so, then why not carry this assumption over to the case of conjunction? When it appears to one that there is a red cube one represents redness and one represents cubeness and one represents some kind of inherence/togetherness relation.

"I think the basic notion is the notion of experiential contents (where these are given by accuracy conditions that reflect the phenomenal character of the experience). Which properties an experience represents depends on which accuracy conditions it has."

I think this will likely be our sticking point. But I go the other way. I don't think we have a basic notion of representation, it is more epistemologicaly mushy than direct awareness of instances of redness and shape and size. I prefer characterizing content as the basic stuff that we are directly aware of during experience; properties that have good credentials. Then I'd like to see whether we can develop a working notion of accuracy out of it and causal relations. I don't think veridicality is basic either, but that our understanding of it is parasitic on our understanding of correspondence and appealing to it tends to bring in assumptions like experience has proposition content before we have yet begun to characterize the content of experience from the ground up.

"My central thesis will be true if experience represents that the cat is weighing down the hammock ."

Yep. All you need is one case and nothing I said above casts doubt on this. I actually think this case is the strongest case and tough. Now, I'm actually inclined to think simultaneous causation is impossible for other reasons, so this case doesn't work for me, although Michael Huemer has a very nice paper defending the view that simultaneous causation is all around.

I'll try to respond to the other comments soon.

[John Morrison]

Hi Susanna,

I have a couple of questions. I hope to have something more substantial to say later, after I've had time to think more about these issues. I apologize if I misinterpret what you've written or overlook answers to my questionss that are already in your paper.

(1) I imagine that some philosopher somewhere holds the view that causal relations are necessarily indeterministic. (If such a person exists, perhaps they think that the referent of the term 'causation' is fixed by our best scientific theory, and our best scientific theories only refer to indeterministic relations between events.) Yet I have an intuition that, if we represent the ball and the lights as bearing a relation, we represent them as bearing a deterministic relation. Do you deny my intuition, the conflicting theory of causation, or that our perceptual experiencs of causation are completely neutral about the metaphysics of causation?

(2) I imagine that some philosopher somewhere holds the view that, while God is the only cause of every event, God sometimes takes motion from one event and transfers it to another event. (I suspect that Malebranche, Spinoza, and Leibniz held views like this.) Suppose that at some point in the future we discover that this view is correct. I don't think that we would subsequently claim that all our previous experiences of, for instance, knives cutting bread were perceptual illusions. Instead, I think we would claim that we were wrong when we described our experiences as representing causation because in fact we were merely representing God's transfer of motion from one object to another. I think that this indicates that the accuracy of our perceptual experiences does not depend on whether there are causal relations between objects, which in turn indicates that we do not perceptually represent causal relations. Perheps our experiences represent some more general kind of relaton of which causation is only one mode, which wouldn't be sufficient for the Causal Thesis given that when we represent something we don't thereby represent all its modes. Do you deny my intuition about what we would say if we discovered that God is the only cause, or do you deny the relevance of such thought experiments to determining the accuracy conditions of our experiences? The latter seems like it could lead to trouble because similar sorts of thought experiments play a pivotal role in many argument for Semantic Externalism.

(3) As you mention, someone who rejected the Causal Thesis might reject it because they think that we only represent unified motions. You challenge such a person to give accuracy conditions for experiences of unified motion. Why is this a fair challenge given that you don't give acccuracy conditions for experiences of causation? Can't they just respond: the experience is accurate when the motions really are unified, and I will leave it up to metaphysicians to give an account of what it is for motions to be unified.

(4) Someone somehwere presumably holds the view that the motion of an object is a *part* of the object, and that that motion is itself composed of lots of little motions, which count as its parts. (One might read Spinoza as accepting such a view, although Spinoza probably wished to identify material objects with motions). That person might also hold the view that in some cases a part of the motion of an object is transferred to another object. Such a person would be able to give accuracy conditions for experiences of unified motion - such an experience is accurate just in case a part of one object is transferred to another. Significantly, just because part of x's motion is transferred to y it does not follow that x causes y; presumably a part of an object can be transfered to another object without the first object causing the transfer (perhaps because it doesn't satisfy necessary counterfactual conditions for causation). what would you make of such a proposal?

Again, thanks for a very interesting paper. If this was a normal conference I'd now say "OK, I'll shut up and let you answer".

[susanna]

Hi John, thanks for the excellent questions. They have a common theme: What is experience committed to when it's committed to there a being a causal relation between two things? I don't say as much as I should about where to draw the line between what experience is neutral on and what it takes a stand on, or about what makes this the case (eg, whether it reflects a supposed ordinary concept of causation, so that in this case the contents of experience are constrained by the contours of an associated concept). This is a shortcoming and your questions bring it into focus.

Re (1), I'm not sure I have the view you're imagining firmly in mind. Would it have the consequence that when the baseball hits the window (or any other putative paradigm case of causation), it is a 50-50 chance whether the window breaks - or are we talking about less dramatic chances? If the dramatic view turned out to be correct, then yes, I think experiences representing that one thing caused another would all be false. I'd have to hear more about the non-dramatic view before I could answer.

Re (2), I agree that the thought-experiment you describe is relevant to deciding whether the causal thesis is true.
I don't quite see the difference between the God-involving view you describe differ and the view that says that the causal relation between the knife and the bread consists in part in an intervention by God. These seem like terminological variants.

It seems reasonable that experience is neutral on the mechanism by which one thing causally produces another. Being neutral on God's transfering motion could be another instance of that. If so, then if we discovered that the theory you describe is true, this would not make any experiences of the knife cutting the bread falsidical.

Re (3) on unity: If all you say is that the experiences represent the events as being unified, then this won't reflect the asymmetry of (many) causal relations. I see the analogy you're drawing between the position you sketch and my position - both try to say, 'don't ask me to tell you much more about what experience is committed to in representing this relation between these events, all it tells you is that they stand in this [causal/unity] relation'. But there are ever so many kinds of unity. It doesn't seem quite right to say, 'experiences represent unity, but is neutral on much of its metaphysical nature', because it is doubtful that 'unity' there is picking out a single relation. In contrast, 'causation' seems to do better. My colleague Ned Hall thinks there are two kinds of causation, but even if he's right there are many more than two kinds of unity.

Re (4) and the composite-motion theory: Cool theory. I guess you're thinking, maybe this is a kind of non-causal unity that experiences could in principle represent, so that in the cases where I say experiences represent causal relations, your proposal says no, all it represents is transfer of motion. What about the stationary cases, eg where the cat is weighing down the hammock, or where a waiter (or a pedestal) is holding up a plate? Does your proposal say that experience represents transfer of motion in those cases? If so, that seems a bit strange since nothing is moving. If not, then (for all you've said so far - maybe there could be more to the story...) it will predict that experience finds nothing in common between the unity relations in these stationary cases and in the dynamic ones. I was thinking of that as a cost, though you might disagree.

Thanks again -

-susanna