Ethical thought experiments show us what we think we ought to think when unencumbered by all the shed load of physiological inputs, prejudices, memories, adrenelin, politics and sociology accompanying an actual moral decision. Since these are the meat of most moral questions I'm not sure what value that could possibly have.

Then there's the bizarre dissonance between on the one hand clearly investigating something we don't currently know whilst on the other presuming we have such flawless knowledge of all the other variables that we can perform this mental trick of holding them steady whilst we conduct our 'controlled trial'.

Beliefs and meaning both refer to things apart from the mind. In fact, they give evidence that we have a mind. — Sam26

I might ask you the same question I asked Banno. If they refer to "things" what are these things, for you? Where are they, what do they consist of? How do they come into existence?

By way of summary, a relation between an agent and a proposition such that the agent holds the proposition to be true is not the same sort of thing as a neural network. — Banno

OK, so can you elaborate further on what kind of thing this relation is? I'm something of a physicalist, so at some point I tend to ask the question of what anything we're discussing consist of, in terms of matter and energy.

If I get an answer in those terms I can build back up from there to a point were it's more useful to talk about the thing (it's impossible to talk about human relations in terms of their atoms - doesn't mean we're not just atoms, just that the relations can't be talked about that way).

If I get an answer in terms of more complicated physics I generally have to take that as fact on trust because I don't understand complicated physics.

If I get the answer that the thing we're talking about exists but not as a physical thing anywhere, I start to think I've probably nothing more to say in that conversation. I don't do God, Realms of Thought, Things which existed outside the universe, Mind Substances, or fairy dust.

But I'm quite happy that other people do, my enquiries are not intended to show these people up (unless they're espousing some position I think is harmful), I just want to know if the conversation has any interest for me.

I have yet to clarify, from your comments so far, what kind of thing you think 'belief' is. A 'relation' you say between a person/cat and a proposition. So if I may just ask for a bit more clarity;

Where is this relation, what kind of thing is it? For me, the 'relation' that houses are bigger than apples is in my brain (we can actually track down and manipulate this relation such that people are baffled when houses don't fit into apples). The proposition "houses are bigger than apples" is also in my brain (language centres, images drawn to the occipital cortex by the hippocampus in response to internally voicing those words...). The relation between that proposition and my actions is in my brain (intention, 3D image manipulation, option weighing all have their respective cortices and networks). But, whilst I understand you don't find it useful to talk of beliefs this way, I'm still a little unclear as to whether that's because you think they exist in some other realm, or whether you have them exist somewhere physical (the brain?) but just don't find it useful to talk about them that way (like talking about politics in terms of the interactions of atoms).

To summarise, are beliefs, for you, something non-physical (in some other realm of existence), or are they physically represented, but just not usefully (or perhaps even possibly) talked about that way?

But to our case, the assertion that for every belief there exists some neural equivalent is in that class. If you say "here is a belief for which there is no neural equivalent", I might reply that there is, it's just that we haven't found it yet. So the proposal is not falsifiable. And yet it is also not provable, because one cannot provide an exhaustive list of beliefs. — Banno

I'm sorry if I'm being slow, but I'm still not following you.

"If you say "here is a belief for which there is no neural equivalent", I might reply that there is, it's just that we haven't found it yet "

- True, but not what's happening here. I'm not trying to argue simply that there might be neurological architecture corresponding to certain beliefs, I'm trying to understand what has made you so convinced that there isn't (remember you didn't say any of what you said conditionally, yours were absolute statements about what could and could not be the case).

if you are going to use the word one way, you can't come back and tell those who use it differently that they are wrong. — Banno

This is quite disingenuous. Look back over our exchange...

I think it goes without saying - despite my having to say it - that there is no particular neural network that in some sense corresponds to or represents my cat's taking it that the floor is solid. — Banno

...is where I first objected. Before that, the most I had said was that a belief can be thought of as a neural network. I didn't declare that beliefs should be thought of as neural networks and then tell anyone using it differently that they are wrong.

Now look at the nature of the statements I've been responding to...

That he takes it to be the case that the floor is solid is not something that is represented in a part of my cat's brain. — Banno

A belief is not an item of mental furniture. — Banno

No! A belief is not a mental state. — Banno

I'm struggling to see me saying that beliefs can be thought of as neural networks (and arguing against statements that they cannot) as being "us[ing] the word one way, [and] come[ing] back and tell those who use it differently that they are wrong.". Yet somehow to read your comments above as being examples of some different approach which I should try to emulate.

How are you not doing exactly what you said shouldn't be done? You're using 'belief a certain way (where it does not have neural correlates) and telling everyone who uses it differently (or more specifically just those who use it the way I do) that they're wrong.

Would you agree with the idea that all we have to work with is our perception of reality... our perception(representation, if you like) of the tree. That seems to be underwriting your position.

Am I mistaken about that? — creativesoul

Not if I've understood you correctly, no. That is basically what I take to be the case. But if discussion of the neurological correlates of belief might be off-topic, then discussion of different approaches to realism certainly is. All this has been laid out in other threads and whilst I don't mind repeating it at all (you never know when something new might turn up) I really don't think here is the right place to do so.

Isn't all this just recognising that words are used in language games? The original doesn't have any greater claim to authenticity does it?

'Believe' in some language games means 'as opposed to know', but in others it means, 'has a tendency to act as if'.

This is usually well within our grasp to handle (by which I mean 'grasp' as in 'attainable', and 'handle' as in 'deal competently with'. Not 'grasp' as in 'hold with the hand' and 'handle' as in 'thing that allows the opening of a door'.)

I suspect that you believe that my house has a front door, and yet had not given that belief any consideration until just now. — Banno

Then in what way would I have such a belief? Are you perhaps saying that because I would picture your house with a front door, I have that belief, even though I haven't yet actually pictured it thus (until you mentioned it)? Doesn't this turn the ascribing of beliefs into a sort of guessing game?

Notwithstanding that issue, even if we were to accept this future-possible behaviour as indication of a belief, isn't that tendency still caused by the state of my neural architecture? Isn't the belief that your house has a front door simply subsumed in the belief that all houses (without prior evidence to the contrary) have front doors. As things stand {Banno's House} is just an example from {all houses without any prior evidence of front-door-lessness}. It seems strange to say I have a belief about your house in favour of saying I would have a belief about your house if I thought about it.

I mentioned before that there seems to me that there is something a bit unfair in sugesting that I ought produce empirical evidence. Look at what you just quoted - and flip it to what you might be arguing - is it that you wish to argue that every belief can be thought of as equivalent to some neural architecture?

Because that's a hair's breadth away from the all-and-some proposition that for every belief there is some equivalent neural architecture.

In virtue of their logical structure, such propositions are neither provable nor disprovable. — Banno

I'm not well versed in things like logical structure, you might have to explain this in a bit more depth. As far as I can tell if you supplied a belief for which I could not provide you with a neurological pathway that is the equivalent of it (produces the same behaviour), wouldn't that be proof? So far you've talked about cats and object solidity and I've outlined (very briefly) the neurological architecture which is the equivalent of this belief. Is there some particular belief you have in mind that you're thinking doesn't have a neurological correspondence?

What I am objecting to is your calling that neurological explanation, in every case, a belief. — Banno

Yep, I hear you. I think we can perhaps shelve that particular part of the disagreement. You think there is a value in distinguishing some neurological architectures which facilitate certain behaviour as corresponding to (some) beliefs, but other neurological structures which (to me) seem to do the same thing, you think are not best thought of as corresponding to a belief. This seems really about the utility of the label which, I'm sure we can both agree, is context dependant. Let's focus on the meat of the disagreement which is that some beliefs do not have neurological equivalents. The opposite (that some neurological behaviour-related pathways might not be called 'beliefs') seems trivial by comparison - unless you think it important that they're literally never referred to that way.

So if there were beliefs that did not directly influence behaviour...? — Banno

What would be an example of such a belief? This may well hinge upon what we each mean by belief. I take it to mean 'a tendency to act as if...'. I have a belief that the pub is at the end of the road means I have a tendency to act as if the pub were at the end of the road. My cat believes the floor is solid mean my cat has a tendency to act as if the floor were solid. By this definition, there could not be a belief which did not influence behaviour. You've used the term 'taking something to be the case', which seems to me basically the roughly same idea, but perhaps yours includes expectation in a way mine doesn't, is that it? Are you saying we can expect things to be the case without ever yet behaving as if they were? If so we've got a whole other are of cognitive processes to go into if you want to find me an example of a belief (in this sense) which doesn't have a neurological equivalent.

If you are going to appeal to an authority you had best reference it. — Banno

You're right, but I'm afraid it's a paper copy only (not a free journal). It's in 'Experimental Brain Research' 1988 vol71 p491-507 and was conducted by Giacomo Rizzolatti.

Has he lost his belief that this surface is solid, or has he lost his belief that any surface is solid? OR have you just "deleted" the concept of solidity....? — Banno

As I said to CS, this is speculative, so there's no hard and fast answer, I just dislike it when things are dismissed summarily as being or not being the case when diligent and painstaking work is actually being undertaken to try and find out. With regards the experimental data I'm aware of (by no means all there is), it is that they have lost the connection between the identification of the object and the motor movements they believe (my definition) are appropriate in response to that object. I don't know that I'm doing a very good job of explaining it, but I'm trying not to turn every post into a cognitive psychology textbook. Yet you, quite reasonably, want firm evidence, so it's difficult. In a few words - there are two pathways visual perception takes, one deals with (among other things) object recognition, the other deals with (among other things) sensorimotor responses. There's a crucial communication between the two as the bit dealing with sensorimotor responses doesn't distinguish objects in memory and so can't predict or manipulate them in 3D in the mind. When you disrupt either pathway the relevant skill is lost, if you disrupt the connection, the wrong sensorimotor responses are applied to the object of perception. To me, that parses as the fact that the monkey has lost it's belief that this object (the one recognised by the ventral pathway), has this physical response (can support my weight - is solid, for example). The monkey has lost it's belief that this floor (the one it's currently looking at and recognising as a floor) is solid (can be walked on in an normal manner).

But, poor monkey! — Banno

Indeed. There is also a study of a poor woman who had part of this happen to her naturally as a consequence of brain lesions!

I would say that. Does the above hold true regardless? — creativesoul

Well, within the F5 region of the motor sensory area in the cerebral cortex there are 85 neural clusters which code for hands and feet responding in different ways to different shapes and densities of the surface they are about to contact with. Disrupting these areas causes the hands (or feet) of monkeys to treat the surface they are about to interact with as if it were of uncertain shape or density. With these regions acting normally, the hands or feet responded to the surface as if it were the shape and density it actually was.

In another study, the visual pathways feeding these areas were tested for attentional variation (to make sure that something different was not going on related to variability in attention) They found that in both 2D and 3D object representations, there was no statistically significant variance in signal between attentional and non-attentional conditions (though slightly more difference in the 2D experiments, which may show 2D images have less significance to the motor regions).

Either way, if you wanted to present an argument that the tendency to treat the floor as if it were solid was not represented somehow in the brain of the animal, you'd have to provide an alternative explanation for the effect of changes to the F5 region in treatment of object shape and density.

This is highly speculative stuff, none of this is set in stone so I think most neuroscientists and cognitive psychologists are open to suggestions. What I'm not sure is so helpful are bare assertions which don't address the evidence we have spent decades painstakingly acquiring.

What role does the tree play in an individual's belief about the tree? The tree is an irrevocable elemental constituent of all belief about trees according to the position I'm advocating for/from. It's one of the elements within the correlation itself. Trees are one part of the correlations drawn between them and other things. Without trees, there can be no belief about them. — creativesoul

The tree is itself a belief about the unity of branches, roots leaves and the disunity of insects, pathogens, bacteria soils, water, gases, nutrients etc (without which the 'tree' would cease to exist). It is a unit that has been created by a belief about how the world is structured. Why are the cells constituting the Xylem part of 'the tree', but the water they carry not? Why, when we look closely at the microscopic fibres of the roots which meet soil water at a molecular scale do we decide the tree ends and the soil water begins. Why are the Mycorrhiza inextricably fused to the cells of the tree's roots not part of the tree, but the fungal cells fused to the algae are part of the Lichen?

I don't know if that's the sort of thing you were asking, but the role 'the tree' plays is as a placeholder of sets of beliefs, but it's no less a belief itself than those which are 'about it'.

The cow/brick metaphor doesn't work for you - perhaps cows and bricks are too similar. Let's try cows and assets. You have a cow that is an asset. I'm pointing out that not all cows are assets, and not all assets are cows.

So for some purposes it does make perfect sense to talk of cows using the term "asset". But not for all. — Banno

No, you're asserting that not all cows are assets. I'm asking you for some evidence or line of reasoning to back up that assertion. You're saying that not all beliefs can be thought of as their equivalent neural architecture (not all cows can be thought of as assets). This is true of cows and assets because there are some cows which are not owned by anyone and as such cannot represent an asset, which must be owned in some sense. Now I want a similar argument for beliefs and neural architecture. Why (and how) can some beliefs exist, but be impossible to think of in terms of neural architecture?

How does the cat (a biological entity - if we're being physicalist) 'take the floor to be solid' without doing so using it's brain in some representative way? I'm not seeing what route you'd like to take here.

If you want to say the solidity of the floor is not represented in its brain you are just flat out wrong about that. Models of the solidity, consistency and constancy of objects are not only located in the brain but we've a very good idea exactly where they are and how they work.

If you want to say the cat applies some general model but does not have a neural representation of this exact bit of floor then again you are just factually wrong about that. We know from scans of savant memory vs normal memory that individuation of environmental models does occur (it's just that non-savants filter it out early on). The cat certainly registers and models that exact piece of floor, even if it's not looking at it.

To continue with your metaphor - I can see, and entirely agree, that it sometimes makes sense to talk of a cow as an asset and sometimes it doesn't, the accountant wants to hear of it as an asset, the veterinarian doesn't. But there the similarity with beliefs and neural architecture ends. For what is true of cows and assets (that some cows aren't assets at all, in any sense) is not true of beliefs and neural architecture. All beliefs which directly influence behaviour can be talked about in terms of their neural architecture, there aren't any exceptions.

The only way out I can see is to include, in the set of beliefs, all negative beliefs. That because I walk without a crash helmet, home from work, I must therefore 'believe' that I am not in the path of a flying object, that I'm not going to be mind-melded by aliens using their remote mind-intervention rays, that I'm not going to be targeted by psychic attack from Russian agents trained in telepathy...

If you seriously want to make the claim that I 'believe' all these things, then I agree that these beliefs do not have neural correlates (they have the absence of neural architecture to that effect). But if so then I cannot see where you're going with such a model.

At what point in the chain of being “existence” or “life” do you think this ability disappears working your way down. — prothero

Somewhere below Primates, Cetaceans and possibly Elephants.

At the moment we know that people do not report self-awareness, nor do they have any memories which result from self awareness if their rostral dorsolateral pontine tegmentum is damaged together with the left ventral, anterior insula and the pregenual anterior cingulate cortex. Every single person so far who's had fMri scanning in vegetative states has has disruption in the network between these three areas.

The anterior cingulate cortex has a lot of specialized neurons called spindle cells, which are found only in primates, cetaceans, and elephants. Thus animals less evolved than these would seem to be lacking a structure in the brain which has been experimentally shown to be required for self-awareness. It seem reasonable to conclude, for now, that this would mean animals who don't have this functional area are not self-aware.

It's possible, of course that other animals simply use different cortices for the purpose, certainly the root of consciousness itself (as in awake, as opposed to knocked-out) is in the rostral dorsolateral pontine tegmentum which is a feature of the brainstem, something shared with all other animals with a brain of any description. This would still exclude insects, however.

part of every perceptual feature is a proposed collection of environmental interventions which are in accord with those goals. To put super special emphasis on this; the goals and environmental interventions are part of perception, and our perceptual features are laced with (summaries of) them. — fdrake

Excellent. This really cannot be emphasised enough.

(Direct realism (feature) ) The properties of a perceptual feature of a perceptual event are identical with those of what the perceptual event is directed towards. — fdrake

OK, so where in this system do priors from past models come in. Some of our perceptual features are delivered by other areas of the brain, often quite unrelated to the object of the perceptual event, occasionally not even sensory data at all. For example the brown table with grey edges I used in one of my earlier posts. One perceptual feature being collected into the summary is the prior model of {tables of that sort} which presents them as being all brown. In our summary we have an all brown table. We've misrepresented the grey edges because our expectation that they would be brown was a stronger model than the weak data that we might have caught a glimpse of a grey edge on this one. But that part of the summary, that perceptual feature, is not identical to the object the perceptual event is directed towards, it's identical to a summary model of previous similar objects.

So what do we say of this input? Is it still 'directed' toward the object of the perception event (on that it is certainly a prediction of it, not a prediction of something else)? I could go with that, but it seems to be stretching the use of 'direct' when compared to your definition of 'indirect'.

Active accounts of perception have exploratory and goal-directed environmental interventions as part of the perception itself. In that regard, hallucinations, or the argument from dreaming, are effectively not forms of perception because two necessary components of perception have been denied of it. Only the relation or absence of relation of descriptive content remains. — fdrake

I think this is a good move as it removes what might otherwise be a distracting tangent.

In a typical instance of perception; not some weird Ramachandran stuff where he's managed to convince the body that the knee is a table by perturbing expectations of the causal structure of the environment — fdrake

You know what I'm going to say to this. Your use of 'typical' may be applicable to philosophical positions (and I accept that's the name of the forum so consider this off-topic) but Ramachandran's work (nor people like Anil Seth or Peggy Series following him) is not leading to the conclusion that the processes in these atypical cases are themselves atypical. The atypical case reveals a step in the typical process which, in the healthy subject, is merely passed over unnoticed.

the interventions we enact are not causally separated from environmental hidden states, even if the inferential summary of environmental properties are. — fdrake

I agree, but the environmental hidden states causing our interventions are not limited to properties of the object of the perception event.

When there is a successful modelling relationship between a perceptual feature and what it regards; or an intervention and our overall model of the causal structure of our environment; the overall perceptual state we're in, and its perceptual features, are indeed informative of our environmental objects. But informational dependence in that sense is not the same thing as saying the properties of the apple in total are existentially dependent upon our perceptions of it. — fdrake

Again, I'd agree with this, but highlight again the extent to which priors are affected by factors outside of the current perceptual event. I think you can make a very convincing argument that our environment in general must directly inform our models of it, that we see 'the world' as it really is (in the sense that properties of the world are the only source of data for our models) no matter how far outside of some particular model's markov blanket such properties are. But... This does not necessarily translate to any given perceptual event.

For indirect realists; the proof of indirectness is inferential representation.
For direct realists; the proof of directness is causal contact. — fdrake

If this is true then both are indeed the case. The disagreement seems to be misguided in those terms.

Perhaps what might be necessary is to posit some implications of either position which would be different if the alternative view were taken. I suspect your position is not as representative of the broad swathe of direct realist positions you include at the top of the post, but I may be wrong. Such an exercise might draw out some of those differences.

Now that's fine, so long as it is clear that this is not the same as using it to refer to the way we take things to be - the folk definition of "belief", if you will. — Banno

OK, but aren't we (the rest if us here) discussing exactly what is meant by 'belief'? Doesn't that very discussion undermine the idea of some clear folk definition from which any other use of the term would be an aberration? I think you perhaps ask too much of language in expecting there to be some use of 'belief' which is a) consistent and b) coherent between uses.

All the term has to do is get the listener to act in the way using it was supposed to do. If I say "I believe you're lying" I expect you to either take offence, admit it, perhaps be less confident about your proposition...some collection of possibilities like that. So long as something like that happens I've used the word successfully. It doesn't have to refer to anything, unless I wanted to you attend to some object by my using it.

So analysis what a term refers to but sticking religiously to all folk uses seems like a search for unicorns.

Nonetheless, I get that a definition such as the one I might use in some specific sub-section of academic work will not be applicable to a sentence like "I have a belief that...". You can't say "I have a [neurological architecture] that..." because what follows is a linguistic statement (describing the state of affairs one act as if were the case), not the actual tendency to act as if. We'd have to say "I have a neurological architecture which causes me to act as if..." before the same statement can follow, which does not have quite the same meaning.

So I understand what you mean when you say that a belief is the (imagined) state of affairs one takes to be the case. But to a physicalist, that definition falls short because we then want to know where such a thing is. Without positing a domain of thought (and I sincerely hope you're not suggesting we do that), we need to know where such a state of affairs is, what does and imagined state of affairs consist of, physically. It's not in the real world outside of our minds - it's not the actual arrangement of such, otherwise beliefs could not be wrong. So what are we physically talking about when we say "an imagined state of affairs"? My answer is a particular arrangement of neural connections, hence that's what a belief is, physically. A belief is 'the imagined state of affairs one acts as if were the case' and all 'imagined states of affairs' are particular neural arrangements in a capable brain. The alternative is dualism.

So to

you can use "cow" to mean brick, but you can't build a house out of cows. — Banno

Really? This is the most important point. I understand the rhetorical flourish, but are we really talking about cows and bricks here? In what cases does referring to beliefs as the physical neural architecture cause anything more than a trivial grammatical stumble compared to talking about them as states of affairs? I think this really comes down to the comments you made which are far more objectionable to me than the insistence on the actual referent for the word 'belief'. Your assertion that there is not even a neurological correlate for beliefs. That the taking of some state of affairs to be the case does not have a neurological correlate at all. If you really believe that then I can see how you think we might be talking about cows and bricks, but if you really believe that then you are either manifestly wrong or your fundamental view of the world is so non-physicalist that we will struggle to communicate, you might as well be talking about fairy dust.

All behaviour is initiated by neurological signals (not necessarily in the brain, mind), for all behaviour which has an expectation (acting as if...) that expectation has to modulate behaviour and the only way that can be done is by other neurological signals. thus any expectation that is actually modulating behaviour is coded somehow in neurological signals.

It's this, I think that best answers @Sam26's question, which is why I brought it up in the first place. If one takes a belief to simply be the state of affairs one acts as if were the case then one's ontology (am I using that right?) becomes overwhelmed with every negative belief. I'm currently acting as if a nuclear bomb is not about to drop on my head, so is that a belief of mine? Do I really have, as a belief, the negation of every single possible state of affairs (except the one I take to be the case) because I'm acting as if they were not the case? That seems inordinately messy to me. Looked at neurologically, however, the problem dissolves. The things I have as a belief are those which do, in reality, modulate my behaviour. any state of affairs which is not physically modulating my behaviour is not a belief I have. adding in the reality of what is happening in the brain (or at least our best guess - it's not an exact science) is the only robust way I can see of pruning the otherwise infinite set of things I apparently 'believe' by negation.

This point "it's trivial to identify who benefits and suffers from past transactions", is asserting that wealthy people have benefited and poor people have been injured? — Tech

No, just that it's more likely than not. A person born wealthy is more likely to remain wealthy. A person born poor is more likely to remain poor, so it is more likely than not that any current wealth disparity is the result of previous wealth disparity than it is the result of something like effort or intelligence.

Governments can only make policies on the basis of averaging (not a case by case basis) so government policy is better assuming wealth disparity is unjust than that it is just.

Free market prices are such that quantity supplied and quantity demand are the same. Real markets are unfree to varying degrees. The distance of real prices from equilibrium prices is a function of unfreeness. Why is your view superior? — Tech

Hopefully I've been explaining that. But in summary it comes down to three factors.

1.Demand has a lower threshold, we have certain basic needs and we can be forced, on pain of death, to pay whatever price we must to obtain them. There is no point at which people will consider bread 'too expensive'.

2. Demand is not a naturally occurring feature. Culture creates demand and pre-existing wealth provides more influence over culture than others.

3. The moment there is even so much as a single law, that will benefit some supply over another, laws affect supply and demand, the wealthy affect laws.

If the seller doesn't turn his property investment into a positive return, then he doesn't make money, so hunger kills him. If the buyer doesn't acquire housing, then exposure kills him. What is the salient difference between buyer and seller position? — Tech

Nothing. I don't think I mentioned a salient difference in that respect. I was simply pointing out that the demand for essential goods cannot be zero and so the suppliers of those essential goods can use this fact to fix the price. At no point will it become 'too expensive' to buy.

Are you pointing out that the lower price is too expensive for some farmer-buyers? — Tech

Yes. That's the point I'm making. If no other landowner is willing to sell for less than £5500 then that is the price land will sell for, it is set by the lowest amount the landowners as a group are willing to sell for. At no point in time will the wouldbe home-owners say "I'm not paying that", they have no choice they have to have shelter, so their demand has nothing whatsoever to do with the price.

Assume a single landlord owns all farmable properties. He decides to sell each acre plot for £0.001. Most farmer-buyers can "afford" land. Unfortunately, the "line" to buy is so long that most farmer-buyers never get land. — Tech

Yes. This would theoretically happen. Farmland itself in that sense is not an essential commodity, I'm talking about essential commodities, property particularly. There is no "line is too long" issues. Absolutely every person needs a house, the line is exactly the population of the country, it may get longer if the landowner sets the price too low (people might decide they want two or three houses), but the opposite is not true, the line will never ever get shorter no matter what price the landowners set. This is the disparity. The landowners, as a group, can set whatever price they like on houses and at no point in time will that figure be pulled down by any 'willingness to pay' restrictions. As such, its reasonable to say that the buyers 'willingness to pay' has no influence on the price.

Would this also work for unconscious beliefs? — frank

I think so, but it's difficult to prove. Experiments with conflicting sensory inputs seem to show that subconscious adjustments are still made on the basis of prediction (the actual experiment is quite complex so you might be best reading the paper directly if you don't want to just take my word for it). I don't think they're in any way conclusive, but together with some very successful models based on heirachical prediction, I think it's reasonable to assume that subconscious prediction is possible.

The problem is, if subconscious prediction is possible then computers can have beliefs, and some people don't like that.

we wouldn't say a tree believes it should grow toward the sun. — frank

Indeed. I think to capture the use of the term you have to limit it to features of (or linguistic representations of features of) a mind. Ramsey expresses beliefs as probabilities and I think that is essential. Probability requires prediction, which a tree can't do (I don't think they can anyway).

Might it be better to think of belief as an explanation of behaviour? Therefore, that the individual holds the (stated) belief is an explanation of the tendency to act. — Luke

Yes. I think somewhere in all this most of the disparate opinions seem to still have a locus around the idea that a belief is related to a tendency to act, so referring to it as the explanation of that tendency seems like it might be uncontroversial.

For me, I prefer to use belief in the same way as say 'injury'. Having an 'injury' consists of the actual physical damage, not the description of it. In the terms you outlined above, having an injury might also explain some tendency to behave a certain way (limping, for example though of course one does not 'have an injury that..' ), but I can also locate the injury, I could even cut out the injury and take it somewhere, it's definitely a physical thing despite also being an explanation. This is how I see beliefs, they are physical structures, I could cut one out and remove it in the same way as an injury.

The reason for this is because in my academic work I had to take account of things like the effect of lesions on behaviour and talking about them affecting beliefs enabled this. We could theoretically create, modify or remove a belief by physical interference with the brain.

I understand, however, that such a definition is not going to be useful everywhere. What I object are efforts to somehow rule it out as incoherent or just wrong (not that I'm suggesting you're making such an argument, I'm just explaining my position).

No, but the statement of a belief is not what causes the tendency to act either, is it? Its the arrangement of neural connections in a certain form.

What I'm trying to draw a parallel between is the idea that a belief can actually be a certain arrangement of neural connections in the same way that a physical law or feature can actually be some arrangement of matter. We have to render both into statements to talk about them, but neither actually consist of the statement.

You seem to be saying that beliefs are necessarily a different kind of thing where the fact that we have to render them into statements carries some additional burden not applicable to physical laws or features. It's this step that I'm not understanding.

So how do you see this differently from, say, laws of physics? They seem to me to have the same property, yet we talk about the relations as consisting of physically manifest patterns, not the statements thereof.

It wasn't so much that we talk about beliefs in terms other than linguistic renderings, just that we don't, in other areas, infer this to mean that they consist of linguistic renderings.

We cannot talk about physical laws other than by their linguistic rendering. The law that "less dense materials rise relative to more dense ones". But we don't say that the law consists of the linguistic rendering, we say that the law consists of some physical relation between the two materials.

The same would be true, I think, of any such description of our models of how the world is, all the relations and patterns we see have to be rendered into statements in order to talk about them, but we also talk about them as if the had some physical existence in the world. I'm just not getting why there'd be any resistance to treating beliefs in the same way.

So you agree that it is a linguistic rendering of an attitude or mental state? — Luke

I certainly think it's a mental state. I think whether it's a linguistic rendering is up to us, I mean it's just a word we can define it how we like, to a point.

The problem I have with restricting the term to statements is it just leaves us wanting of a term for 'that which causes a tendency to act as if something were the case' when it is not rendered as a statement.

In binocular rivalry experiments, for example, actions can be generated in response to a combined image set despite the subject only being aware of the dominant image. How are we to talk about such a tendency toward action if the word belief is reserved for that which is rendered as a statement?

Maybe we could distinguish the two types of tendency to act, but I'm not sure I see either the advantage or the precident. Other states of mind (such as affect) are described in terms of the state itself, not the linguistic rendering of the state.

Too much about the perceiver and not enough about the perceived (or about the relation). I mean, it's not "arbitrary", as you said it was (uncharitable perhaps). — jamalrob

Yeah, I meant arbitrary relative to the world. @fdrake has made the same point to me before and he's right, I tend to ignore (in my simplification) the interaction caused by the fact that part of 'the world' is our thinking about it. So yes, I am guilty of perhaps focusing too heavily on arbitrariness.

What I mean by it is that we should be mindful of the effect our reification of objects has on their properties. It's important, I think, in issues like perception, physiological sensation issues, even the dreaded 'free will' debate (what is it that is free and free from what).

The way I see it, this is just a truism. Maybe you're interpreting it more strongly. — jamalrob

Does my reply to Marchesk help any?

How are real objects dependent on our models of them without it being anti-realist? — Marchesk

It's to do with what 'an object' is defined as. Imagine the world consists of just an heterogeneous soup. That's all there is, one object. Any object we define out of that soup is determined entirely by the arbitrary line we draw distinguish it. So whether something is black, or whether it's black-and-white depends entirely on where we draw the line around it. All of its properties depend on what we determine 'it' to be.

As I asked, if that's not a belief then what is? — Luke

Then we're possibly on the same page. I asked the same question about the assertion that a belief is not a mental state. If it's not a statement, not a mental state, not state of some dualistic realm, then it seems to me that we're running out of things it could be.

I think jamalrob is arguing that how an object looks, tastes, feels only applies to perception. There's no such thing as what an object looks like without someone seeing it. The indirect realist goes wrong by assuming there is, and then proposing the additional mental intermediary. But there's no need for the intermediary if the act of seeing is what something looks like.

If that sort of argument works, then the debate is rendered moot. There's still a realist question of what objects are independent of perception, but they aren't like perceptions. — Marchesk


Aye I think that's roughly where I stand. — jamalrob

Really? That didn't come across at all from what I read (not your fault, I'm not well versed in all of the philosophical background).

If that's the case I think we probably agree. I'm in favour of what I think is called 'model-dependent realism'. The idea that there is a real world out there, but the objects in it and their properties are dependent on our models of them. Basically, as I've discussed at length here some time ago, I don't think there's support for even so much as drawing a line to mark the end of one object and the beginning of the next in the world 'out there'.

So how an object looks, tastes, feels only applies to perception. There's no such thing as what an object looks like without someone seeing it. Sums up how I see things too. Only I'd add that the very delineation of one object from another falls into the same realm.

Funny how difficult it can be sometimes to see common ground. Thanks @Marchesk for pointing the way.

I'm interested now if you think there's anything I've said here which doesn't fit into that view, or is my error only in mistaking your position.

So now you're making the argument you just told me was a straw man. — unenlightened

How so? The assertion is only tangential, I was trying to show how it's important in some fields to get a grasp of the effect different stages in the perception process have on our final concious image of an object/scene. If it's straw-manning your position then you can either explain how or we can just drop the entire angle. It's got nothing to do with the rest of my post, which you've not yet responded to.

From what I recall of similar arguments in the past, the conversation always faltered over the meaning of "direct" and "realism". It would inevitably run aground on semantic disputes. — Marchesk

I see. Possibly the solution is to focus on some practical consequence of either view, some behavioural reflection of the different beliefs.

What exactly is modified? — jamalrob

The data from the sensory organs relating to the object as it is in the world.

you mean the perception is modified. — jamalrob

I think, for the sake of clarity, it's best if we all stick to 'perception' as encompassing the entire process, so no, I don't mean the perception is modified. I mean the perception constitutes an amalgamation of two (or more) data sources, only one of which is the actual object in the world.

The perception is the result of, or is constituted by, modifications of light, electrical impulses, and so on, but that doesn't say anything about a modification of perception or experience as such. — jamalrob

Yes, but it is also the result of prior expectation which may well have absolutely nothing to do with the object under consideration.

Let me try an example. Say I'm looking at a table which is larger than my field of vision. I only glance at it. To anyone who inspects it closely, it is brown with a grey border. The table I see (the one I describe to others, plan to eat off, recall later etc) is completely brown, even at the edges. This browness is not a property of the table in the real world. It's a property of previous tables of similar sorts. My vision at the periphery does not register colour, my brief glance wasn't enough to take in the edges. But the table I have in mind (the one I describe to others, plan to eat off, recall later etc) does not just have no edges, or blank edges. It very clearly has brown edges. It has the edges I was expecting to see, not the ones generated by the wavelength of light from the actual table.

We can call this a 'direct' representation of the table (just a mistaken one), but if we do I'm not sure what 'indirect' would mean in this context.

Therefore, beliefs are not pre-linguistic or non-linguistic. Unless a belief is something else? — Luke

But isn't this just begging the question? If a belief is "an attitude to the world (or a mental state if you like) when rendered as a statement." then it obviously follows that it must be linguistic, but this is no more than to say "if a belief is linguistic, then it is linguistic "

I think what @Sam26 is asking (though I'm not so sure now) is whether this need be the case. Obviously if we simply declare that beliefs are statements, then they are de facto linguistic, but that doesn't answer the question of whether declaring beliefs to be statements covers everything we need beliefs to do.

think you underestimate the subtly of what a stick user can detect, and the vibration of stick on book, stick on concrete, and stick on stone are very different. — unenlightened

That may well be the case, but it would only stand as a counter argument if you're suggesting that it would be literally impossible to pull of such a ruse. That the stick was so subtle as to be able to detect any and all changes. If not, then all you've done is shown that I chose a bad example.

Well if the blind man or the bat could pass through barriers that I could not or vice versa, that would be evidence that we were detecting different worlds, possibly. — unenlightened

Who said anything about detecting different worlds? Detecting a representation of the world is not the same as detecting a different world. All its saying is that the concept or image that you have is not a faithful copy of the world, it is modified by factors unrelated to the current state of affairs.

at least some of the time, that is exactly the argument; that we cannot tell the difference between a dream and reality. — unenlightened

Schizophrenics cannot tell the difference. Unless you're excluding them from 'we' then there is such a case. Which means examining how we make such distinctions is a worthwhile endeavour.

Is the blind man's perception direct or not? Is it direct if he uses his hands with no stick? This is where I want an answer. Is touch direct perception? At what length of stick or fingernail does it become indirect? — unenlightened

At the point where the data is not faithfully translated. My eyes, ears, fingers, the blind man's stick, all just send signals to the relevant cortices of the brain - filtered, but direct. We are completely unaware of their doing so, we have no concious connection to that process. Within those cortices the signals are combined with backward-acting signals from parts of the brain completely unconnected to the data received from the object. This combination of two (or more data streams initiates action (still without any concious awareness) - it might be a tilt of the head, a shifting of the gaze, a move of the fingers. This provides more information (filtered by the expectations from the backward-acting neural connections), which is in turn subject to modifications by parts of the brain unconnected from the actual object. Finally after what might be several hundred iterations of this process we gain a concious image or concept of the object (we can name it, imagine interactions with it etc).

The process has several hundred opportunities for information not related to the object being sensed to alter the final concept we become conciously aware of (the one we name, plan with, talk about etc). How is that not 'indirect'?

What in general intervenes between world and experiencer to make experience indirect? — unenlightened

Hopefully I've answered that question above.

the question then is how that representation is perceived. — unenlightened

With the part of the brain responsible for receiving signals from the various sensory cortices. If you're really interested in the heirachical brain model I can go into it, but I sense perhaps not.

perceive the representation, or do I merely perceive a representation of the representation? — unenlightened

No, you directly form your concept of the image from the various cortices involved in the analysis one level below. In the sequence A>B>C>D, D does not receive signals direct from A, but it does from C. It's not a complicated system, we don't throw our hands in the air and say "well if D does not receive direct signals from A then I suppose nothing receives direct signals from anything".

Why is it less problematic to perceive a representation of the world than to perceive the world? — unenlightened

It's not about how problematic it is. There is a point prior to our becoming aware of the properties of an object where data unrelated to that object (in the world outside our minds) can alter the data we originally received from it. That's basically as close to a certainty as we're going to get in neuroscience at the moment, it's not really up for debate at the moment. So if the object we all talk about, plan with, name, recall, etc is formed from both signals from the actual object and signals from elsewhere in our brain unrelated to the object, then our model of the object must be indirect. It is not directly related to the object but related to both the object and our prior expectations of it. It is an amalgamation of the two, and so not a direct representation.

No — Banno

Really? That's the quality of discussion you want? You just espouse some dogmatic opinion and any contrary view is silenced with the standard cliche that it's "too wrong to even respond to". It's not the Solomon-like crushing blow you think it is, it's just boring.

A brick is not a cow because if I asked any for a cow to build a house with I would not get what I want. If I asked for a brick to milk I would not get what I want.

A brick is not a cow because no one uses the words 'brick' and 'cow' in that way. That's not the kind of case we have here. If that were the case you would not have to be telling us how to use the words correctly, there would be no discussion.

On classes, a brick is not a cow because one is an inert clay cuboid for building and the other is a type of animal from which we can extract milk.

It's not difficult. So humour me...

What is your evidence that there is no neural net corresponding to the belief that the floor is solid?

I think, I hope, i don't have to be claiming that the blind man's world is made of stick vibrations. Merely that the world he detects is the world, and not a representation of the world. — unenlightened

If someone took up the kerb and replaced it with a pile of books of similar size and shape, the blind man's conception of the object he detected would be indistinguishable from the concrete kerb. His prior expectations have influenced his conception of the world. He expects there to be a kerb there, he's received no signals to the contrary so he conceives of a kerb there. If you asked him to 'read' the object at his feet he would not proceed to do so, he would simply presume you were mad. So how can he be detecting 'the world'? He has a concrete kerb where a pile of books is.

You might argue that his model is still derived from 'the world', but that would just be straw-manning the indirect realist argument. No one is saying that perception is not initiated by signals from the outside world.

If "the world he detects is the world, and not a representation of the world" what would the counterfactual be. Take your claim to be false, what would be the case to show that it was false? What would a 'representation of the world' be like that differs so much from how things in minds actually are?

would anyone demand that to be direct, the table-top would have to project a rectangular shape on to the retina? Is there actually a naive position that is somehow corrected by the idea that perception happens from a perspective and in a certain way? — jamalrob

I think this is perhaps the crux of the problem, and the point that @fdrake was working towards. If the direct realist is going to say "we don't literally mean 'direct' as in the images on the retina are the same as the objects in the world", then the indirect realist objection is somewhat undermined. But it works the other way too as the indirect realist can say "we don't mean 'indirect' as in there's literally no relation between the perceived objects and those in the real world" and so the direct realist objection is undermined.

But to have a perception of an object which is modified in some way (and even fabricated to some extent) from the real-world source of the sensations which precipitated the perception, is most definitely 'indirect'.

There is (without a shadow of doubt now) some other influence on the image/conception we consciously have of the object other than the sensations we originally received from it. Our prior knowledge influences the image we create and our prior knowledge is not restricted to being about 'that object'. It might be about others like it, or even unrelated to the object at all.

That is, as far as I can tell, basically the definition of indirect - took a route with significant nodes, didn't go from A straight to B, had some turning points along the way...

So what exactly is the distinction the indirect realist is supposed to be making that the direct realist wants to deny? That's what I'm still not getting here?

I'm just disincline to think that must be some neural network that corresponds to each and every possible belief, stated or unstated. — Banno

Yes, I'm gathering that very strongly from your repeated dogmatic dismissal of the idea. What I was actually asking for was any evidence or reasoning whatsoever to support that belief. Do you have some neuroscience demonstrating it to be the case? Do you have some idea that there isn't room in the brain for all that data? Do you just really really really not want it to be the case?...

what you said was that "a belief can be a particular neural network", which might be more indicative of the development of a technical sense of "belief" in neuroscience. — Banno

I said it 'can' be because it is yet unproven to be. we can't measure individual neuron activity yet, only clusters of activity. since we can't measure individual neurons we can't see what's going on at the scale required to code for such specific beliefs as "this bit of this particular floor is solid at this moment in time". Hence I try to still refer to it (when I remember) as a possibility.

We have about 86 billion neurons. Since each neuron can have as many as 10,000 synapses, that's 8.6 x10^13 architectural elements. Are you telling me you can think of more than 8.6x10^13 different beliefs that you and your body require at any one time, because even if you exercised a unique individual one of those beliefs every microsecond you could still go nearly three years without repeating one.

I am not making the claim that some particular neural network could not also be described in terms of some belief. — Banno

I can't make sense of this. How could a particular neural network be described in terms of some belief if

there is no neural net corresponding to the belief that the floor is solid — Banno

Are you suggesting that there are neural correlates for some beliefs but not others? That, in some respects, would seem even more odd than what I originally thought you were claiming. What would distinguish those which 'make the grade' and get coded as neural architecture from those which don't? and again, where would the others be?

Think it through. Are you going to argue that there is a particular neural net of some sort for the floor's being solid? Another for the cup being in the cupboard? Another for there being a poppy in the front garden and another for that poppy being pink? One for each of the innumerable unstated things that are taken as true as I get up to open the window? — Banno

Yes. I asked you for sources, you do realise you're espousing a theory about how the brain works (or in this case, doesn't) you're own incredulity does not constitute evidence.

Let'ssstart with what you think is wrong with there being a neural network for each of these beliefs. Do you think the brain will run out of space, if so how much space do you think a belief such as the ones you list takes up, and how much space do you think we have. You must have some pretty clear ideas on this for you to be so certain there isn't room. Or is there some other reason you find the concept so incredulous.

Secondly, if the belief (the tendency to act as if) is not stored in the brain, where do you think it is? In the air, the ether, some dualistic field?

One for every conceivable belief that might be stated? — Banno

One for every 'conceivable' belief? With what do we do the conceiving if there's not room in our brains for all the conceivable beliefs?

So the next observation might be that, because we can be mistaken, what we take to be the case is distinct from what is the case. — Banno

How can this be so. The belief that the cat is in kitchen has merely been replaced with one that she is in the hall. That too might be mistaken. They are not two different kinds of thing ("what we take to be the case" and "what is the case") they are two instances of the same type of thing ("what we take to be the case" in response to some evidence, and "what we take to be the case" in response to some new evidence). I'm not seeing what the concept of "what is the case" is usefully doing here.

there is no particular neural network that in some sense corresponds to or represents my cat's taking it that the floor is solid. — Banno

Where are you getting that idea from? Could you cite me the source(s) you're working from on this.

How do you explain, for example, how some patients with damage to the parietal lobe, are unable to reach accurately towards visual targets that they unequivocally report seeing, if there's no neural network corresponding to "taking its spatiotemporal position to be the case" then why does damage to some network destroy that particular belief (as demonstrated by behaviour)?

I'm not saying there isn't something going on in the mind. I'm only saying that we don't point to things in the mind to defend the idea that we have beliefs. No more than we point to something in the mind to define a word. There are things that occur in the world that reflect these things. — Sam26

OK, that makes perfect sense, but you asked about the 'form' of a belief and that still confuses me. I understand your three ways in which a belief can be expressed (although I prefer to just call them all actions and have done with it), but I'm not seeing what type of answer you want to the question of 'form' that wouldn't simply be 'neural architecture'. The form of a belief is neural architecture, it's identified (in the absence of being able to read that architecture directly) by behaviour. Acting as if some state of affairs were the case indicates that we have some particular neural architecture which is responsible for our tendency to do so. Is there still something missing from that description?