Scientific Representation: A Dilemma for Realism

[@KaneB]

Scientific realism: kzbin.info/www/bejne/l2G4qXmegLScfMk

[@yassirel653]

Pleas kane i wanted to ask you about something you said in previous vedios : My life doesn't worth living my question is , what is the reason that you can give it to yourself to continue living if the life doesnt worth living

[@marsglorious]

The Lion has returned. Stylish coat. Exploring a new look. Hitting new highs. ❤❤

[@KaneB]

Thanks dawg!

[@lhyere9730]

Excellent video! I definitely agree that it is better to think of representation as a triadic relationship between agent, source and target. However, I'm not fully grasping why this 'individualistic' view undermines our understanding of science as a collective effort. Conversation is a collective endeavor, but at any time what any word or phrase in an utterance stands for seems to be fixed by the individual's intent. What enables us to communicate, I take it, is that since we belong to shared linguistic community, we have picked up certain habits as far as what words/phrases are taken to 'stand for' in English. I take it that the collective nature of science can be explicated in similar terms, i.e. practitioners in a field can develop coherent discussions about models because they have picked up certain representational habits by engaging with said field.

[@CognitiveOffense]

I've been claiming, boldly unresearched and unqualified to make such pronouncements, for a while that all communication is strongly restricted to Abduction. I think this video is another way of coming at the same problem and I find myself very sympathetic to its conclusions. What I mean when I say "All premises are statistical.", what someone else interprets it to mean, and what a broader community interprets it to mean are all related in some way, but not identical. In physics, I liken it to the entropy vs complexity. Entropy obviates the history of the universe to say "this is a random distribution of gases in a box constituted from interchangeable molecules" and there are many conclusions you can make about what happened before or will happen next, with that model. However, that presupposes a reductionist view, whereas an understanding of complexity includes the hereditary information of particular molecules in a box. Meanings of language follows similar heritable histories. We mere mortals can't use that level of complexity in daily life, so we obviate to an "entropic" gist .The source and target can each change non-linearly in the real world, but to prevent chaos overriding any ability to communicate, we assume the statistically likelihood we suspect has the highest chance of being the intended reference and run with that assumption. Consider the technical jargon of subject matter experts for heritability and equivocation examples of language. How words were used influenced how words are used which influences how words will be used, but this is a system where feedback loops and strange attractors may exist. But this is just an argument I think probably makes sense... probably... depending on whether you, dear reader, use words the same way I probably do. Thanks for another great video.

[@atha5469]

👍

[@octopusgoat2502]

Could we get around the dilemma if we were to say that groups of scientists can come to an agreement about what they take the models to represent? It seems that if an individual is able to specify that they take the fmri to represent Patient X's brain, they could also reach an agreement with a group of their colleagues to the same effect.

[@KaneB]

Yes, I take it that specific groups could explicitly agree on the target of a particular theory, model, or image. I'm not sure whether this solves the dilemma, though. There are contexts where we seem to talk about theories or models as products of science that can be shared by "the scientific community" as a whole, and that can develop over long periods of time -- I can describe, say, the history of the modern evolutionary synthesis and the evidence that supports it, where this isn't simply a matter of what is produced by a specific research group.

[@johnhugon67]

If two scientific theories make exactly the same predictions then the theories are equivalent. In the same vein if multiple different "sources" produce the exact same representation then the sources are equivlent. In the FMRI, as you stated an individual can view the proton precession/emr/human brain as the source. But isnt there then there an issue in trying to force one "true" source. The image is a reperesentation of all of those things because all of those things are the exact same with regards to the specific reperesenation. Why is there a need to narrow down to a single target in the first place?

[@KaneB]

I'm not sure that there is a need to do that, but if we don't do that, it's not so obvious in what sense the image can be taken as an accurate representation. Let's say I take an fMRI of the brain of a schizophrenic person. If I present the image as a representation of a patterns of blood flow in a particular person's brain, then we might say that this is idealized but provides an approximately accurate picture; it shows us something close to the truth. If I take the present as a representation of "the schizophrenic brain", as exemplifying certain general features of the brains of schizophrenic people, it might be completely misleading. Perhaps this particular schizophrenic person has unique characteristics that make them unsuitable as an exemplar. Is the image accurate or inaccurate? Well, without specifying the target, it looks like we can't say.

[@Opposite271]

@@KaneB I could imagine the realist arguing that accuracy is multidimensional and that it can be accurate in relation to one source while inaccurate in relation to another. This could imply that a model has multiple fuzzy truth values.

[@KommentarSpaltenKrieger]

Very interesting. In my understanding, the target of the representation will be defined by the scientific community, not by individuals. In the case of brain scans, the target of the representation is neural activity, because this is what the neuroscientific community "agreed" to want have represented. This can only change if the neuro-scientific community decides to change the target. As for the problem that the target of representation depends on individualistic psychological factors (i.e. on people wanting to have different things represented by the same source), I would argue that the professional attitude of those who use models will include a shared "representational" aim. I.e.: All neuroscientists collectively, but also individually (qua being neuroscientists) will have the fixed target of representation of "neural activity" when it comes to fMRI scans, because this is what they are interested in professionally. While it is possible for them to want another target of representation (for instance proton decay), this would not make sense for them. If they were physicists, the case could be different. But physicists don't care about fMRI scans at all. Actually, only neuroscientists, neurologists, psychiatrists and a few others care about fMRI scans - for all of them, I'd argue, given shared professional attitudes, the model will represent the same target. So, the model means the same for all who use it, even though it could mean different things to different people. Is this an anti-realist perspective? I think it is, but one with strong realist flavours. After all, there is a universally shared target of representation among all the "model users" if I'm correct.

[@KaneB]

I think that, if it's sensible to talk about the aims and beliefs of the "neuroscientific community" as a whole, then we can use that to fix representational targets, and so avoid the dilemma. I'd agree that, when we think about science, it's sometimes useful to think of different fields of science as unified entities, where we're abstracting away from the features of the individual people involved. I would tend to take such talk as highly idealized, and in that respect anti-realist.

[@williamanon2050]

I could be misunderstanding/just not thinking about this well, but isn't this a solution: What if we are also "statement realists" (I'm sure there's probably a better term for this lol). What if we say that logical statements, models, etc., are things that actually exist or at least can be thought about independent of someone making the statement or believing the model. We could then say that in this nebulous realm of statements about the universe, some more accurately capture the actual nature of the objective universe. Then we get around each individual person potentially having different targets for models or differing conceptions of scientific models, because we just say that whatever those conceptions or models are, some more accurately reflect objective reality than others.

[@KaneB]

Hmm, I'm not sure how this avoids the individualism horn of the dilemma. We can grant that some representations more accurately reflect objectively reality. The question is, when we consider some product of science, say an fMRI, what exactly is the representation? Taken as a representation of patterns of proton processions, it may be highly accurate; as a representation of a given person's brain, it may be fairly accurate but highly idealized; as a representation of "the schizophrenic brain" in the abstract, it may be extremely misleading. Now if some individual scientist presents this as a representation of a specific brain, we might judge that this representation accurately reflects objective reality. But we can't say that the image itself accurately reflects objective reality, because the image only has that representational content *as interpreted by this individual scientist*.

[@davidantinucci8027]

I suspect that the likeness between the way most humans generate representation in their human mind (and can agree on the target of their act of representing) does the work here - i.e. the targets of human's representations are sufficiently shareable by their shared cognitive tools. A sufficiently different cognitive process of representing the same state of affairs by a different kind of individuals, like say bats, would make a representation become "lost in translation", but that's more a function of cognitive tools, not of the entire idea of representation.

[@gert8439]

Right. Most humans will agree the sky is blue, because that's how a normally functioning human experiential toolkit works. That's the third person falsifiability of publically accessible (observable/measurable physical stuff) which science relies on. But different species with different experiential toolkits will create different types of representations, like the bat as you say. And as the conscious experience itself is private, we can't be certain that 'what it is like' for you to see blue isn't 'what it is like' for me to see green. But as long as we use the same signifiers consistently, it will still work. Further, if colour is 'created' by minds/brains, then even if we both see a blue sky, it is still a representation. And that goes for all sens6ry exerence.

[@InventiveHarvest]

Meaning of models and such changes with context - it depends on the domain of discourse. Most typically, the author of a paper will have words in it to explain the context the model applies to. People will have different interpretations of a model, but they will still understand it. But I guess for people that think the word "cat" is indeterminate, I guess they will also view models the same way.

[@hian]

Yeah. I don't think there's a real dilemma here. If anything it seems to just reflect what you'd expect under scientific realism/naturalism. For example, the word "cat" is just a string of words and letters. It doesn't inherently refer to felines in a subject-independent sense, yet it functions perfectly fine in communication between large groups of people due to linguistic consensus. Pointing out that an individual could use "cat" in a novel way that would render communication difficult or impossible has no bearing on whether or not the creature being referred to under the standard definition exists or not, or on whether or not functional communication is possible using the standard definition. Similarly, if you were to use the terms in a theory to mean things that would render the theory unintelligible or impotent, that bears no relevance to whether or not the theory is accurate using the definitions that it was constructed with in mind to describe/explain/analyze/predict a particular phenomenon. There's no dilemma, just a mistaking the map for the territory fallacy described as such. The function of the theory and its accuracy is predicated on a particular determinate use of language which requires linguistic consensus to be understood. Lacking that consensus, or worse, ignoring that consensus; does not render the theory inaccurate. All it does, is presentational obfuscation of the fact that the words of the theory no longer describe the theory. I.E it's no longer the same theory. It's just a string of words equivalent in appearance to the theory, but in substance something entirely different. The easiest way of demonstrating this is to just consider a statement like, "Yesterday, I saw a cat." If I change the definition of each word to mean, "Today, you heard a dog", then the two starements are still equivalent in appearance even though the words are different in substance. Thus, while "I saw a cat" would no longer mean just that, making it an unaccurate statement in so far as personal observations of cats one day prior is concerned, it doesn't really have any implications for the accuracy of the statement when used with its original definitions; or any other combinations of words whose meanings are equivalent to the original definition.

[@KaneB]

@@hian >> Similarly, if you were to use the terms in a theory to mean things that would render the theory unintelligible or impotent, that bears no relevance to whether or not the theory is accurate using the definitions that it was constructed with in mind I agree, but that has nothing to do with the argument of the video. The dilemma is: First horn: We can take theories and models as representational, but then we have to think of these as items in the minds of particular individuals, and lots of people (including lots of realists) often prefer to avoid this "individualist" perspective. Second horn: We cannot specify any particular target of the representations, which is to say that there is just no fact of the matter what definitions the theory/model was constructed with, and so no fact of the matter whether it is accurate. >> it doesn't really have any implications for the accuracy of the statement when used with its original definitions Suppose that I said the sentence, "yesterday, I saw a cat," but that there was just no fact of the matter whether this meant "yesterday, I saw a cat" or "today, I saw a dog." Presumably, that would have implications for the accuracy of the statement.

[@hian]

@@KaneB I'd say it has everything to do with the argument, because the scientific realistic position is that the models employ language in reference to specific and concrete phenomenon which is entirely separate from the issue of whether the language used in the model is understood by secondary parties to refer to the phenomenon of the person who formulated the theory. You can be epistemically skeptic in regards to whether any actual phenomenon exists for which to refer, but it's nonsense to be skeptical of that simply on the basis that swapping meaning around in relation to symbols, obfuscates what phenomenon words purport to refer to. Just because two individuals could use the same words in different respects has no bearing on the issue of whether "a phenomenon" to which the words were intended to refer to on an individual account de facto refer to an actual phenomenon. And, if a scientific realist would argue otherwise, that would be both silly and unwise. If there is a phenomenon, and we decide that a term refers to it, the accuracy of a model containing it is predicated on the consensus adoption of the definitions for the terms. If this isn't done, then the model isn't made any "less accurate". It just isn't the same model anymore because non of the terms refer to the phenomenon to which the model was meant to serve. Again, the model is predicated on a shared appreciation of language, but to say the model is "no longer accurate" if there's no shared appreciation of the language is literally just a mistaking the map for the territory fallacy, since the substance of the model is in the relationships of the phenomenon described within it, not the strings of letters and sounds with which it is communicated. If I use "cat" to refer to felines in a particular sense, your awareness of the particular respect in which I used it is simply not relevant to whether or not there is actually "a cat" to refer to in that respect. Thus, both horns are wrong here insofar as: 1. an individual rejecting or being unaware of the specific point of reference that another individual posits for a term(in a model or not) is not in any way relevant to whether or not a phenomenon exists to which the term referred. Hence no realist should be inclined to see this as an issue unless they're really bad at being realists. Some might, but just because some might does not mean there's an actual issue here. 2. Just because I cannot specificy what a word refers to in a universal sense on other people's behalf by simply using the word itself, it does not follow that there is no fact of the matter for which the word was intended to refer. Nor does the necessity to provide context to inform specificity in regards to reference imply that there is no actual point of reference. There just is nothing in the statement you've used here where the one follows from the other. If you wanted to argue that there is no fact of the matter to distinguish "I saw a cat yesterday" from "You saw a dog today", you could make that argument, but that argument is not made simply on the basis of the fact that people may operate with different definitions. So, summa summarum, Functional language is a product of social consensus. A shared understanding and appreciation of words referring to equivalent phenomenon. That is to say, "I saw a cat yesterday", is functional because there's a shared agreement what the constituent words refer to. The necessity of this for the term to functionally communicate anything is separate from any conversation or argument about whether or not the term accurately refers to "something". The "cat" is(presumably) the fact of the matter, and the word's function an extention of social consensum that this particular combination of letters and words refer to the phenomenon of cats. Now, if you want to be skeptical of whether or not we're even perceiving the same thing in the first place(I.E common anti-realist objections to whether or not it's epistemically possible to ascertain whether people are de facto using words to refer to equivalent phenomenon) then that's entirely fine. However, such objections exist entirely separately from this supposed "dilemma" and do not require this approach in the first place. It's obviously the case, as would most realists openly admit I imagine, that you and I could both be using the term "green" to describe two different phenomenon unawares. However, that issue would be true, again, wholly separately from the question of whether there is an underlying fact of the matter as far as the phenomemon of green is concerned. Nor would the presence of this problem imply a dilemma of language as far as accuracy of models are concerned. If I posit a model like Pythagoras formula for the calculation of a sides of a triangle, and you use "triangle" to refer to squares, it's not the case that Pythagoras formula ceased to be accurate. You are just no longer using Pythagoras formula. The fact that recognizing this requires communication is not an issue for realists. If anything, it's expected on a realist's view as it's not clear what it would mean for terms like triangles to self-contain a line to their reference points. After all, on the realist's view, a word is just a collection of sounds and letters that only acquire meaning through social consensus. There are no platonic forms. There is no meaning as attributes of objects separate from the brains of people.

[@KaneB]

@@hian I read the first three paragraphs of this and as far as I could tell, they had nothing to do with my argument, so I stopped reading there. If you want to rephrase your basic objection in a shorter comment, or if you want to specify the relevant paragraphs of this comment, I can respond.

[@martinbennett2228]

@@hian You have replied to this better than I could have done. In watching the video (thank you Kane) I immediately thought this is about how language works (arbitrary signs and signified referents) more than anything else, so I do not really see a dilemma. I do not think there is a coherent subjective account of language, but interpretation depends on the reality status of the referents and signified concepts. The choice of using fMRI as an example is interesting because it does highlight the distinction between data (which science assumes to be objectively real) and interpretation. Clearly in the case of fMRI there is often likely to be a large measure of uncertainty of the extent to which the interpretation corresponds to reality. Nevertheless those involved in fMRI are unlikely to question whether there is a reality which they are trying to account for, but they would also be aware of the possibility of artefacts that might put an interpretation to be wide of the mark.

[@italogiardina8183]

Representation and reality seem interdependent through mathematics as a meta language, where science emerges as a plethora of truth conditions that compete as social power that hedges as individual influence to gain status in their field of expertise.

[@ostihpem]

Sounds a lot like a case for relativism: if as representation r is true depends on who agrees on what to be represented as r. But one could - at least in theory - work out ways to agree on that. The deeper problem seems for me that the representation r is true iff r‘ (= that what is represented by r). But we will find that we are always only refer one representation to another - like r‘ is a representation like r, just on a supposedly different level, refering to r - so we will never get truth but only relative truth, i.e. if r‘ is true then r is true. Skepticism seems unavoidable here as well, relativism or not.

[@NelsonGuedes]

It is relative to our perception because we don't have access to the thing itself in the world. What we get is not truth but some level of understanding, because the truth itself is not accessible to us.

[@mathusalen1]

What's your opinion of Massimi's Perspectival realism?

[@illusion5342]

There is a fundamental barrier between objective reality, our subjective experience of that reality and our linguistic propositions describing that reality. So we have to work within those limits. Science is the best answer, because it gives the most certainty. I just wonder, what is your exact normative claim at the end? You didn't say it explicitly, but I felt you were implying that we shouldn't believe in science?

[@KaneB]

I don't make any normative claims. I've presented a challenge to scientific realism, but it doesn't bother me if people are scientific realists. I wouldn't claim that you ought not to be a scientific realist, or that scientific realism is irrational, or anything along those lines.

[@unhingedconnoisseur164]

@@KaneB out of curiosity, how would one understand it being a challenge / problem in non normative terms?

[@masterlikesmargarita]

Some questions: What do you mean "it could represent radio signals"?! 'What i write clearly represents letters' - for real?! The image is produced via radio waves and a specific measurement of them. I make a painting with a brush - do i represent the brush in the painting? Clearly i could have this in mind when doing so - doing some sort of metarepresentation. But it must be intended - and i could only do so because it has been institutionalised. The details dont matter you say and go on to argue in such vagueness that it must be 'in some sense'... If you want to know what it represents you must know the details of how it becomes. Also you identify the notion of examples - noone will say that a single fmri image represents 'the brain'; they would only do so in the means of it being examplary. Yes the educated ones will say that an fmri shows you certain functions or movements IN the brain and wont say like a layman may be inclined to that it is how a brain looks like. And an artist may use the image to symbolically represent scientism or reductionism or something idk. But just because you may use it differently doesnt change that it will and can show activities of a particular brain. "It would never have occurred to me to think the sentence had such an aura if I had not thought of how one might say it differently-as a quotation, as a joke, as practice in elocution, and so on. And then all at once I wanted to say, then all at once it seemed to me, that I must after all have meant the words somehow specially; differently, that is, from in those other cases." (PI §607) What is being shown is being shown: ask the person using it what it is used for and stop atomizing everything. It is just as you then go on to say - only thing is we dont need to get to konsensus: one can only sensefully do such because it is konsensical. Because any "intention is embedded in its situation, in human customs and institutions." (PI §337) Insofar collectively representing the same thing is not an issue - rather if there is anything else beyond our constructions we represent. And there i would argue that it is only because of an external world how we can come to form certain constructions or more profoundly even begin doing so... And just as an add-on. You form a strange case for such images; they are not there to give a perfect or accurate representation of the brain; the purpose of an fmri is to show us certain things happening in the brain by being able to measure differences from certain fluids etc (as you see i dont know the details either). And the proper question here is if it does that what is intended accurately. And for on that note one last quote from Wittgenstein: "It is as if one said: 'The clock tells us the time. What time is, is not yet settled. And as for what one tells the time for-that doesn't [belong] here.'" (PI 363) (I dont know if Anscombe necessairely translated Wittgenstein the best; i notice some differences to the german version.)

[@KaneB]

>> What do you mean "it could represent radio signals"?! I mean that patterns of radio signals could be taken as the target of the representation. If somebody were to point at the image, and say e.g. "here there are more intense radio signals" or whatever, we would probably take this as a correct description. It might be useful in the context of learning how fMRI works, for instance. >> The details dont matter you say If the argument for indeterminacy of representation works, it works regardless of how precisely fMRI images are created. There might be parts of the story I told that are simplified or idealized. The key point, however, is that there are complex causal chains involved in the construction of an fMRI image, and the image can plausibly be taken to represent various elements in these causal chains (among other things). >> noone will say that a single fmri image represents 'the brain' It's fairly common to take these images as representing "the brain", in the abstract. Or something like "the human brain" or "the schizophrenic brain". I'm not sure why you deny this. Perhaps you could elaborate? >> But just because you may use it differently doesnt change that it will and can show activities of a particular brain Yes of course, but why is this relevant? The fMRI image is constructed from scans of a single brain. There are many ways of taking the image as a representation. The fact that we can take it as a more abstract representation, say a representation of "the schizophrenic brain", does not prevent us from taking it as a representation of a particular brain (the particular brain that was scanned in the construction of the image). Obviously our interpretation of it as a representation doesn't alter the causal processes in the image construction. But my claim here is that there is an indeterminacy of representation, not an indeterminacy of causal processes. (In fact, I do think there is also an indeterminacy of causal processes; I favour a kind of antirealist, projectivist view of causality. But that's not really relevant to the argument of the video. Everything I say in this video is compatible with causal realism.)

[@masterlikesmargarita]

@@KaneB "we would probably take it as a real description" That is what i say - if someone uses it as such (when he knows that it can be used in such way) then (in this context which is understandable to everyone) it would be a true statement. But there can be many applications and reals to be described (why must it only be able to show one?); there are many legitimate ussages. The thing is via an fmri most want to show something via radio signals and their interaction with other parts. "why you would deny that" Yea this might have been a bit too radical of an expression. But i think i elaborated when i continued by saying that it will be used to give an example. There isnt a 'general brain' which they refer to - to generalize one is not in need to refer to an overarching idea of a brain. But as said - someone can and yes some do; but we will see when we see how they use it. There is context and they will elaborate. Obviously if we would take a picture like this in a vacuum it wouldnt be able to determine its meaning, and for this i would even go further and say that there wouldnt be determanacy for it to even refer. But we arent in a vacuum. (On another note tho: it is interesting to think of it in more levels - like in a sense wether without cultural or imidiat context one wouldnt be able to determine wether it was used as an artistic expression etc; but the interesting fact would be wether or not one could determine that this is a scan of a brain... but there you see again: it will be a scan of it. Being something does not imply refering to it... as little we refer to the great attributions of wood when we make furniture out of it.) To make a last point here: noone wouldnt be able to specify when they would refer to 'the brain'. And they could also say that technically it is only an example etc. I think i answered your questions. The only position where it might seem indeterminal is taking it out of its context. If someone knows the details and what they use it for - they will know what they represent. They can represent something only by the means of that but obviously this logical connection only goes this one way. One uses an fmri to represent x; but only because one uses an fmri doesnt imply this person tries to represent x. (It is sort of the other way around than classifications: if something is a dog, then it is also an animal, and then it is also living... But just because something is an animal it doesnt mean for it to be a dog.) (Edit: the problem i have is that you make it that IT, the picture, the fmri by itself represents. And it does not. We do with it. Because it is produced in such and such ways that it can show this and that.)

[@rath60]

There seems to be two branches to your argument first the data branch and second the model branch. Lets start with the data branch. The image of the fmri is a set of dark and light regions on a medium. Our model tell us that these regions correlate to the body being imaged in a certain way. The human mind has the capacity to learn the shapes and structures of objects and can tell if they have seen a brain or cross-section of a brain that in fact the two shapes match. And so on. But the model is the interesting part. As a student I was taught how to interpret data. This skill creates in my mind a sort of representation of the world that coincides with the representation of other scientist while we do since. I can in fact do science in non-scientific settings. For instance I may strive for reproducibility in baking. Or experiment with separating technique from superstition. Models are then way about thinking of the universe that we attempt to teach and reproduce in other humans. They are memes. Highly infectious memes. Because they inter link. My model of weights and temperatures can be tested in the realm of reproducibility in baking. My model of pitch and vibrational sense can be tested in the construction or application of instruments. Etc. But, given biological urges such as curiosity we see begin to value the doing of science with out notion of its applicability. We see which new memes and ideas what can create for no greater sake than the fulfillment of boredom. Or for fame or recognition or any of the many other mundane reason people have for there actions.

[@ReubsWalsh]

I think maybe, *if* you're wrong, it would be because you're combining experimental or observational descriptions - the data, basically - which have difficult-to-nail-down 'targets' when you treat them as representation - with theoretical models. When a scientist proposes a model - let's say, for a really simple example, that Broca's area is involved in speech - they're using inductive reasoning from data - e.g., fMRI images of people speaking versus not speaking, stroke victims with lesions in that area whose speech became impaired. Then try to prove it wrong. Deliver transcranial magnetic stimulation to the Broca area of a willing volunteer who is motivated to continue speaking. I'm not sure if human neurosciences really constitute a "mature" science, though. We can't integrate across levels of analysis well enough yet. Physicists just need to reconcile GR and quantum. Human neuroscience has to reconcile across social, intentional/mental, biological, and probably electrochemical levels, people are asking is there a quantum aspect too... so at the moment, I think a lot of the field is underdetermined to individualism *and* non-representation, in a way, because a lot of the time you really have to work with a lot of stipulative, and still possibly quite vague definitions, and what you're defining is less a theory about how the world (or, the brain) is, and more a theory about how we could get closer to beginning to properly understand it. To me anyway, it feels like a lot of scientific theory papers are mostly not claiming to describe reality so much as to say "look, here's a pattern that could be important..." In the end, all human communication is underdetermined, there can always be misunderstanding, and no photograph would look qualitatively similar to its target when you zoom in far enough. Maybe I've misunderstood something. I tend to take a critical realist view of science. Like, it's real, but we've got "flatworld" type obliviousness to most of everything, we evolved a way to think rationally within the confines of what humans can affect or be affected by, and it's imperfect and starts to break down at the extreme edges of those confines, plus it's finite, whereas the target ruleset could potentially be infinite (assuming we want to include complex dynamic systems within the bounds of scientific study). edit: your example with the description of yersinia pestis and its proteases - I definitely think this kind of statement is literally true. The unobservable neurons are absolutely literally real, as real as observables like the nose on your face. There might be mistakes in some specifics, almost certainly, but the 'central dogma'(s) of the life sciences aren't 'just a good metaphor' like caloric was for heat energy. They might be Newtonian gravity rather than GR. Psychological and psychiatric unobservables are often explicitly metaphorical, a way of describing a pattern.

[@MMurine]

@Prodigious147 OP's comment was very intelligible to me. Some of the particular scientific language was beyond me, and their meaning was too particular to rearticulate simply or concisely without considerable effort, but it was clearly considered and specific in its use of language. This one might just be on you.

[@MMurine]

@Prodigious147 I anticipate that you have a very narrow conception of utility.

[@MMurine]

@Prodigious147 To denigrate the utility of philosophy for not being essential to 'actual science research' is to blame a football for not helping you catch fish. The rest of your comments are mostly shallow peacocking.

[@MMurine]

@Prodigious147 Check your presentation for solecisms, and ask your psych friend for an evaluation. I think they're onto something.

[@MMurine]

@Prodigious147 When I press this button, he generates three replies.

[@rsm3t]

If the first horn is correct, then the implication is that the languages we use (including spoken/written word, math and logic, illustration, etc.) aren't doing their job. I don't think that is the case, though. Certainly this video carried an idea across to a number of viewers, for example (whether or not those viewers agree with the idea). Of course there can be failures of language, but I doubt such pathological cases are exemplary of your dilemma.

[@Gabriel-sn6yg]

I fail to see how the photo of Frank Zappa is different from the MRI. Somebody could use a photo of Frank Zappa to represent a non-specific guy with a big mustache or guys with big mustache in general... Using an element of a set (or its representation) in order to represent the set is often done even outside of science. Somebody could, for exemple, use the photo of someone in an hair salon in order to represent an haircut they would like to have... It is no different than asking if the MRI represent a particular brain the brains in general. Like MRI, a photo is created via certain mecanisms. One could say that a photo represent photons of certain wavelenght, or the possible energy level that electrons cannot have the same way than one can ask if the MRI represent oxygenated blood or radio wave. And the one about big science, there was a time before numerical photo camera when most people, after taking a photograph, had the photographic film developped by someone else who may very well not not what the photograph represented and I don't think anyone believed that it meant that the photograph failed to represent what it represented, likewise, sometime people will have photograph taken of them by strangers. I believe that demonstrates that something can properly represent something even if not everybody partaking in the creation of the source know what the source represents...

[@KaneB]

>> I fail to see how the photo of Frank Zappa is different from the MRI I agree, it isn't. The indeterminacy of representation arises for representation in general, not just scientific representation.

[@Gabriel-sn6yg]

@@KaneB Ok, thank you. I somewhat missunderstood that part of the video.

[@benmustermann2045]

I enjoyed the video, thanks for making the point. But do you think that representation is just constituted by a source, target pair? And furthermore do you think that we can be arbitrarily liberal assigning Source to Target relations? In my experience there are criterion of fit/fidelity for source target pairings and those tend to be sensitive, not just towards So a group of scientists will, in virtue of their being embedded in the scientific community largely share the same criterion for what makes a valid S T pairing in the context of research.

[@KaneB]

>> But do you think that representation is just constituted by a source, target pair? No; it's more usefully thought of as a three-place relation: Agent A interprets source S as representing target T. So there are agents, sources, and targets. Then there are some tricky problems that arise from the fact that we often represent non-existent things. >> And furthermore do you think that we can be arbitrarily liberal assigning Source to Target relations? Yes. In principle, anything can represent anything else.

[@benmustermann2045]

Thanks for the reply. I agree in thinking that scientific representation is most definitely at least a triadic relation. If you had to flesh it out what exactly do you mean by non existent, possible but not actual, abstract or mistaken, such as the ether? Interestingly though I think that we don’t have to go as far as thinking about non existent objects to arrive at interesting questions about how to think of sources and targets. For instance it might be completely unclear what aspect of the target system is meant to be the actual target of the representation. Furthermore do you believe that successful representations need a different treatment than ordinary representations and are you as liberal about successful representations, meaning that anything can be used to successfully represent something else, without at least an relation of mutual information? And I take successful to mean used by scientists in this context.

[@lhyere9730]

@@KaneB I disagree with this. It is true that you can represent anything as a simple by anything else as a simple. For instance, a salt shaker I can take to represent Kenya if I so desire. However, when you are not doing simple-simple pairings, but want to represent aspects of composite systems, this is no longer true. As an example, you cannot represent 3 entities in a triadic relation by two entities, one of which instantiates a property and the other one of which doesn't - that's just not a coherent mapping. Similarly, if you want to represent a car crash by two bottles of water coming into contact, clearly you must take one bottle of water to represent a car and the other bottle another car, you couldn't take one of the bottles of water to represent love and still say that the two bottles together represent a car crash. If the representation is not merely stipulative, but tries to in some way map the structure of source to that of the target, there are clearly constraints to what can represent what.

[@DeadEndFrog]

yo kane, are you and unknown knowns still attempting to make a collab theory together? Or did i just dream about it?

[@KaneB]

We've recorded loads of stuff... I dunno if or when it will ever be released.

[@DeadEndFrog]

@@KaneB

[@MatthewMcVeagh]

I wonder then Kane, what your positive position is? Are you a different kind of realist, and what should we believe about the products of science?

[@OBGynKenobi]

This is the same as Megritte's Ceci n'est pas une pipe , or that video you made about whether a painting of a landscape is really a representation of the landscape or the subjective interpretation of the landscape by the painter. Furthermore, is this video a representation of you, or when I see it, do I see something different than how you actually looked or acted?

[@cunjoz]

I get the feeling that professor Dave would hate you lmao. I got on his nerve simply by pointing out that dark matter and energy are hypotheses that serve to explain observations and that non-falsifiability is built into some models of dark matter and energy. most science communicators are unfortunately very unfamiliar with philosophy of science and epistemology

[@NelsonGuedes]

That only seems to be a problem if you take representations to be perfect copies of the thing in the world that they represent. If I say the word "dog", you will have some concept of a dog in your mind. Like a picture. But the picture of a dog in your mind will not be the same as the picture of a dog in mine. We might imagine completely different breeds of dogs. Does that mean communication was not successful? Or that the word "dog" doesn't really represent something I the world? I don't think so. As long as there is a general concept of what the representation is supposed to represent, the fact that each individual have their own interpretation of that representation doesn't deny the realism of the representation. Of course ambiguity may arise in communication, but that's ok. It can be disambiguated through further communication. But the ambiguity also plays a useful role in science. Since we don't all have exaxtly the same representations in mind, throufg miscommunication we may uncover a different characteristic of the representation which had not occurred to us before. And since anyone can make a "mistake" having an imperfect representation, their representation might be closer to the thing in the world that we are trying to represent, thus moving science forward. I think this ambiguity of representation is not a bug of science, but a feature which science shares with evolution.

[@OBGynKenobi]

But representations are only valid if we all agree as a culture what they mean. Individual interpretation doesn't count unless it agrees with the wider understanding. This is why science only works if others corroborate an individual's interpretation.

[@KaneB]

I'm inclined to think that even individual interpretation does not fix a determine target for representations. So the problem is deeper than I suggested in the video. That said, I'm not sure though why an individual interpretation doesn't "count" unless it conforms to other people's interpretations. What if I come up with a novel interpretation? Think about art criticism: what's interesting about reading commentaries on artworks is when people offer new insights, new interpretations, that differ from the conventional understandings. Or what if I'm the first person to produce a theory or model? Since nobody else is aware of it, there are no other interpretations available.

[@OBGynKenobi]

@@KaneB yes, but if no one else is aware of it then it only exists as a solipsistic concept. It only becomes a thing when everyone is aware of it and believes that it's true, or false. For example, when Einstein was developing his relativity, only he knew about it, thus it did not exist but for one person. But when he published it and it was corroborated and proved by experiment then it was manifest. Therefore it became a thing when the wider culture agreed on what it was or wasn't.

[@alduin2000]

Why is indeterminacy of representation even a problem for realism? You say that for realism to stay with representationalism, it has to endorse a very individualist view of science so that the representation becomes determinate, but I don't understand why science can't just accept indeterminacy in representation head on (indeterminacy in representation does not entail indeterminacy in reality itself after all). In other words, why can't realism say that the fmri image represents *all* of the things you mentioned simultaneously and realistically?

[@Caylynmillard]

Has philosophy taught you about the importance of vitamin d? Apparently not.

[@philbelanger2]

I don't think this is a problem for scientific realism because you don't need to believe that representations are determinate in order to accept realism. For example, take biology. Suppose that we accept that it is indeterminate what biological models represent. This doesn't change the fact that biologists share many beliefs about their field of study. Most of them, for example, believe that an environment selects for the fittest organisms. A scientific realist will say that we should accept what biologists believe about biology. He doesn't need to take a stance on what models represent.

[@KaneB]

As I see it, scientific realism requires more than merely accepting scientific theories and models. The problem is that there are many ways to "accept" something. For the realist, to accept a theory is to believe that the theory is true. For the constructive empiricist, to accept a theory is to take it to be empirically adequate, i.e. true of the observable phenomena, but there is no commitment to truth in general. Or I might accept a theory as a merely instrumentally useful tool: not true in any domain, but more powerful in some sense than the alternatives. I'm not sure what it would be to say that a model is true or accurate, without taking any stance on what the model represents. What is truth, without representation? Consider: (G) God exists. Is (G) true? Well, the natural follow-up question is: What do you mean by "God"? Suppose I say that God is the omnipotent, omniscient creator of the universe. Then maybe you judge (G) to be false. Suppose I am a pantheist, and I say that "God" is just a term I use for all of nature, and "God exists" is a way of expressing an attitude of veneration towards the universe. Well, in that case, (G) might well be true; at least, there's nothing too controversial about saying that there universe exists. But then suppose I cannot or will not elaborate on my meaning -- I insist that God exists, and that's that, there's nothing further to say and nothing to be clarified. In that case, you'd surely just be puzzled about how to respond. Perhaps I'm talking in some artistic or poetic way. It's not clear that it even makes sense to say that my claim is true or false.

[@philbelanger2]

@@KaneB I would say that a scientific realist is someone who believes the theories that scientists believe to be true. This is different from believing that the models used by scientists are true, because models are often idealizations. Since they usually are idealizations, it is not clear to me that they represent anything. If we interpret them literally, they are, strictly speaking, false. But scientists can nonetheless use models to make inferences. For example, a model can show us what we would observe in an idealized situation. A scientist can then use the model to argue that a set of observations provides evidence for one or more of the assumptions of the model. If the evidence is strong enough, we can then believe that these assumptions are true, even if we believe the model as a whole to be false.

[@philbelanger2]

@Prodigious147 Can you clarify your comment? What isn't a belief?

[@OBGynKenobi]

Grass is only green to humans. It looks quite different to a dog or a bird. Wouldn't it he more accurate to say grass is grass? Because qualities, like color are physically subjective, but the object, grass, is the thing.

[@KaneB]

The specific example isn't so important... if you deny that "grass is green" is true, just choose literally any other statement that you think is true. Its truth will be dependent on the meanings of the terms. So with "grass is grass", if the first use of "grass" means the stuff that's growing in my garden, and the second use of "grass" means the stuff that my brother smokes on the weekend, then "grass is grass" will be false.

[@OBGynKenobi]

@@KaneB but meaning is derived from context. So if the context is lawns then we know grass is lawn grass. Actually would be a good subject for a video if you haven't already... Where does meaning come from?

[@martinbennett2228]

@@KaneB Yes, 'grass is green' was not the best example and not a scientific statement. Science might state that grass has low absorbance between wavelengths of 500 and 600 nm. Your point is about language, but I do not think there is much of a problem with the concept of wavelengths; is there a difficulty in that participants have to recognise this referent?

[@KaneB]

@@martinbennett2228 The point of the example was literally just to illustrate that the truth-value of a statement depends on the meanings of the terms. I don't see any problem with the example from that point of view. Yes, I do think that representational indeterminacy occurs for scientific statements and models. I gave an argument for this in the video, which may be completely wrong, but that is my view. Particular individuals have no difficulty specifying a referent. That's why it's a dilemma: individualism or non-representationalism.

[@Liliquan]

First, probably.

[@ipilotaneva2586]

kane is first

[@mikechristian-vn1le]

This guy moves around way way too much, his head, his body, his hands

[@Doctor.T.46]

As a scientist and philosopher can I please offer a view from both perspectives. What you have described is a 'dilemma' for philosophy, not for science. I've noticed over the years that philosophers have, it seems, become obsessed with science, a subject they no little understanding of, while scientists don't give a 'hoot' what philosophers say... they just get on with their work. Sorry to disappoint my fellow philosophers.

[@KaneB]

Well yeah, scientific realism is a philosophical position. I doubt this dilemma would be of any interest to anybody outside of philosophy, but I'm not even pretending to talk to them. It doesn't disappoint me that scientists usually "just get on with their work" without reflecting on the philosophical implications. If I had wanted to talk to scientists, I would have gone into science, not philosophy.

[@Gabriel-sn6yg]

It is a dilemma for "scientific realism", which is a philosophical position, not a scientific one. A scientist can easily do science without being a scientific realist...

[@Doctor.T.46]

@KaneB You make a few points Kane that I would love to respond to. First, can I say that as a philosopher I love your channel. Your first paragraph confuses me. You are missing an important word, 'context'. All medical imaging has a context, which you appear not to give an explanation of. Scientific realism is based on context...like all science. Your second paragraph also confuses me even more. Why don't you want to speak to scientists? You are expressing an opinion about science, yet you are not aiming for scientists to be involved, limiting contribution to those who, like you, don't understand the topic they are talking about. I've spent all my philosophy career involved in the debate surrounding consciousness, listening to philosophers who have no knowledge of Neuroscience trying to persuade me of the conceivability of zombies, so I'm not surprised that you don't feel the need to at least enquire of a scientist what they think about the topic. I don't mind that you want to keep your discussion amongst philosophers, but imagine how you would react if I misquoted or misrepresented Plato. I'm sorry Kane but on this topic you are misrepresenting science and thinking that by saying that you are aiming it at philosophers and not scientists makes it OK... it doesn't, wrong is wrong.

[@Doctor.T.46]

@@Gabriel-sn6yg Not sure what point you're trying to make. A philosopher can do philosophy without doing scientific realism, so what's your point.

[@KaneB]

​@@Doctor.T.46 >> You are missing an important word, 'context'. All medical imaging has a context, which you appear not to give an explanation of. Scientific realism is based on context...like all science. I'm not sure what you're referring to here or what this has to do with my argument. But you never used the word "context" in your original comment. >> Why don't you want to speak to scientists? You are expressing an opinion about science, yet you are not aiming for scientists to be involved, limiting contribution to those who, like you, don't understand the topic they are talking about. I'm addressing a debate within philosophy of science, not science. Obviously scientists are welcome to contribute, but they're not the primary audience, and most of them are just ignorant of the relevant issues. Most scientists aren't aware of, and don't care much about, the contemporary realist/antirealist debate or philosophical issues to do with representation. >> I'm sorry Kane but on this topic you are misrepresenting science What specifically have I misrepresented?