If you think about what you're saying, then you also agree with me. If something appears or happens that has no prior reason for its existence, its a first cause. Notice the title says 'a' not 'the' first cause. There is no reason preventing our universe from having multiple first causes in the past, the present, or the future. A first cause has no reason why it should or should not happen. It simply does. — Philosophim

But that's stretching the meaning of "first" to the point of vacuity, for the concept of "first" is only meaningful in relation to a recognizable order with a distinguished bottom element. In the absence of a well-defined order, the concept makes little sense, especially considering that a rejection of the causal order doesn't entail that postulated "first" causes can't have explanations in terms of other causes, but only that such explanations are incomplete, vague, relative, ever changing, etc.

5. Infinitely prior, and infinitely looped causality, all have one final question of causality that needs answering. "Why would it be that there exists an infinite prior or infinitely looped causality in existence? These two terms will be combined into one, "Infinite causality.

6. If there exists an X which explains the reason why any infinite causality exists, then its not truly infinite causality, as it is something outside of the infinite causality chain. That X then becomes another Y with the same 3 plausibilities of prior causality. Therefore, the existence of a prior causality is actually an Alpha, or first cause. — Philosophim

You need to clearly distinguish spatio-temporal causality from your murkier concept of meta-causality.

In a similar fashion, Stephen Hawking once proposed a causally closed cosmological model of the universe , in which the universe was hypothesized to be finite but without a spatio-temporal boundary. Nevertheless, he famously asked "what breathes fire into the equations?". But this philosophical question as it stands cannot be translated into the spatio-temporal language of physics. Furthermore, there isn't a consensus that Hawkings philosophical question is even meaningful, let alone how it should be solved or dissolved if it is.

Another possibility you are overlooking, is the possibility that the very existence of the past and its historical content might not transcend the ever-changing state of the present. In which case, the past is open and indeterminate like the future and there isn't a universal causal order.

This has nothing to do with theological assertions jgill. Forget God. It floors me that I cannot get through to other atheists on this. Truly their fear of this being theological terrifies them to the point of being unable to think about it. I am an atheist. I wrote this. This is about base matter. Its very simple. Don't let fear prevent you from understanding it. — Philosophim

What makes you think that you can conceive of a first cause?

I can for example, conceive of, and indeed witness, a pencil line that has a beginning, and I can also start counting up from zero. But these so-called "first events" that occur in ordinary experience are only conceivable to me because i am able to witness or conceive of other events in time and space that occur earlier ...

In my experience of fellow atheists, they often harbor a peculiarly theological belief in "nothingness", in that they seem to reify the notion as a sort of anti-substance that they envisage as existing before and after substance, out of which they construct myths such as universe as having an objective "beginning",or of personal experience as having a subjective end. (They will deny these charges of course, in the usual spirit of "true believers"). But if we reject this ontological interpretation of nothingness as being nonsensical, then how else are we supposed to conceive of absolutely first (and last) events?

None of this shit is "tangible". "Infinite" is not tangible. That's the issue, because it's not tangible, mathematicians are free to create all sorts of axioms which do not relate to anything physical. But when the mathematics gets applied there is a very real issue of the intangible aspects of reality. And if the axioms which deal with the intangible in mathematics do not properly represent the real intangible, the product is "the unintelligible". — Metaphysician Undercover

To be clearer, I meant that an infinitesimal is "tangible" if it can be finitely described as a total computable function,which implies that the tangeable infinitesimals correspond to an undecidable countable subset of the natural numbers.

But note that by definition, an infinitesimal only has to satisfy the condition that whenever it is multiplied by a number of arbitrary large size, the product is always less than some finite constant. This condition can be satisfied purely by mapping the natural numbers onto a data-structure other than a line. So there exists semantics for infinitesimals (and their reciprocals) that does not imply the existence of infinite time, space or information (which is the unfortunate result of misinterpreting such numbers as literally denoting limitless extensions)

This is what happens when we approach the issue of "the first cause". The calculus turns the first cause into a limit on tangible causation, rather than treating the first cause as an actual cause. But if there is an actual intangible first cause then the mathematical representation renders that first cause as unintelligible, being outside the limit of causation, according to the conventions for applying the mathematics. — Metaphysician Undercover

Similarly, the information implied by a limit is relative to one's method of counting. E.g if we define a number n to be greater than every natural number (which we have the right to do), then infinite extension isn't implied if we choose to start counting within a finite distance from n.

The classical theory of real numbers interprets 1.000... and 0.999... as referring to the same equivalence class of different Cauchy sequences. So it isn't necessarily true that the system of real numbers conflates the sequences 0.999.... 1.00..., for the truth of that hypothesis is decided by assumptions concerning the existence and construction of Cauchy sequences prior to their identification as real numbers. For example, a computational interpretation will identify cauchy sequences with total computable functions, whose Cauchy limits might not necessarily be decidable, and even if they can be proved to exist, their limits might not be decidably different or indifferent. On the other hand, intuitionism interprets the meaning of 0.9999.... extensionally as referring to an unfinished sequence of data, in which case the very notion of a sequence, cauchy or otherwise, as having a definite limit is denied as absurd, meaning that not only is 0.9999 distinguished from 1.000..., it is also distinguished from any other instance of 0.999....

Perhaps we ought to say that the Real numbers cannot be interpreted as directly referring to Cauchy sequences, unlike in the case of the Hyperreals, on pain of the Cauchy sequence interpretation being in conflict with the Archemedian property of the reals that it's axiomatization imposes by fiat, but which the Hyperreals sacrifices for the sake of an illusion of creating "more" numbers.

Also, lets be wary of non-constructive interpretations of Hyperreals, for otherwise one ends up having infinitesimals by fiat that do not denote anything tangible. If we stick to constructive principles, then contrary to popular belief there cannot be more hyperreals than natural numbers, let alone of real numbers, meaning that hyperreals are just reorderings of the naturals , but whose operations aren't necessarily recursive.

Unless a clear, non-debatable physical example arises the things uncaused may be the empty set. — jgill

A resource-conscious set-theory that only expresses transformations between existent sets, could in principle be developed by introducing "negative" sets, such that the empty set denotes the union of equal sets of opposite polarity, whereby the resulting set-theory operates in an analogous fashion to the string-diagrams of particle physics in which energy is purely transformative without being created or destroyed.

But for some reason the traditions of logic and set theory have remained entrenched in structures such as toposes that forbid an initial object from having incoming arrows from other objects, i.e. their initial objects are strictly initial, which to a layman leads to the unnecessary impression of logical origination.

Not sure what you're saying here but definitely a fan of Kripkenstein. — Apustimelogist

"According to functionalists, mental states are identified by what they do rather than by what they are made of" - IEP

Yet functionalism leaves the very nature of "doing" unspecified, so it is hard to think of what functionalism rules in versus out. The concept of functions/doing is part descriptive and part normative, and related to metaphysical presuppositions about the nature of time, causality and counterfactuals. For instance, can "doing" purely consist of synchronised motions like the contents of a movie? or is agency, causation and the notion of counterfactuals involved?

Kripke came to mind for similar reasons, in his astonishment to learn that the meaning of mathematical functions is intensional, i.e implicit and normative, as opposed extensional i.e explicit and descriptive.

(Data might be interpreted as expressing a function, but cannot ground the meaning of the function or make the meaning of the function explicit, since the latter's meaning is inexhaustible from the perspective of the user of the function who understands it normatively, while being open to interpretation from the perspective of an observer of the purported function who understands it descriptively)

For me, meaning is functional. If our behavior is functionally explained by brains entirely then meaning is as well. — Apustimelogist

If functions are regarded to be nothing more than tools, then it would seem that the intensional meaning of functions is entirely dependent upon the intentional state of the investigator who applies them.

It seems to me that the identification of meaning and function per-se doesn't distinguish function realism from function anti-realism and idealism. (Kripkean skepticism comes to mind again)

The philosophy of mind (which in spite of appearances isn't a particular subject but concerns the whole of the subject of philosophy) is part of science, in so far that the purpose of science is considered to be explanatory in the sense intended to satisfy the existential questions of a particular human being.

The techniques of science and even it's formalized theories can be considered instrumental, but if the purpose of science isn't considered to be instrumentally pragmatic but explanatory in the above sense, then there exists a semantic or explanatory gap between the tools of science and it's supposed goals, which must be filled somehow, leading us back to philosophy and it's patchwork of vague and apparently inconsistent pre-theories

So if we reject the idea that science and philosophy have distinct goals, then what you have described under the heading of the philosophy of mind, is the pitiful state of science as a whole. Also your summary of AI is interesting, because it reflects society's recent obsession with Machine Learning that has up until recently, ignored the normative discipline of symbolic reasoning, which must be addressed if AI is to scale to more difficult problems in a fashion that is reliable and understandable, but that direction opens the can of worms known as the Philosophy of Language, which is at the heart of Philosophy of Mind...

As i see it, the mind-body problem is but one example of the semantic under-determination of scientific theories, and one's tolerance for semantic under-determination depends on what ones goals are.

I think you are giving idealism a realist interpretation, by interpreting " the mind" as a speculated theoretical object or posit, with your infinite-regress arguments resembling those used to attack indirect realism. Ironically, Berkeley's arguments against representationalist materialism were that he found it to be incoherent for reasons which are very similar to yours.

There is no "mind" posited in Berkeley's arguments for subjective idealism in the literal sense you assume, but only ideas referring to the thoughts and observations of the individual.

Nevertheless, Berkeley apparently remained uncommitted to the solipsism which many consider subjective idealism to imply, for although his arguments for idealism were based only on ideas, he was apparently open-minded with regards to the truth of the rationalist doctrines of causality and the external world. Like Malebranche and Hume, Berkeley didn't consider causality to be reducible to observations, for he understood observations in themselves to be inert, like the video frames of a movie. So if causality and externality were to exist, he argued that they must exist in some other mind that exists apart from one's ideas, namely in the mind of god, which ironically leads back to realism.

(I consider Berkeley to have shown that realism is ultimately a theological notion - the speculated existence of external reality in physicalism doesn't seem any less theological to me than Berkeley's mind of god)

This is not an acceptable explanation of causation. An assignment of causation does not exclude the possibility of other things having the same effect. So in the example above, saying that heat causes water to boil does not exclude the possibility that something else as well, such as a drop in pressure, could also cause water to boil. That A is judged to cause B does not exclude the possibility that something else might also cause B as well. — Metaphysician Undercover

You've misunderstood me. Yes, there can be multiple causes for an effect, but when testing for the existence of a causal relation in a series of repeated trials that check that consequents of type B allows follow after antecedents of type A, then it must be assumed as a working hypothesis that there are no other possible causes of B other than events of type A. For otherwise a successful test might only indicate correlation between As and Bs.

You presumably agree that each video frame of a movie isn't the cause of the next video frame in the movie. So even if video frames of type A are seen to always occur before video frames of type B, such that they are in perfect correlation, then you would not want to identify that relation as causation. No?

Which is the reason why counterfactuals come into play. For causation isn't supposed to merely refer to perfect correlation. At least, that isn't how the concept of causation is used by the sciences, in which causes refer to conditional propositions in which the output of the conditional is a function of the input.

I've never seen the concept of causation described as being an interpretation of counterfactual logic. I've always seen it described as the product of inductive reasoning. You know "causation" extends to ancient Greece, and was discussed extensively by Aristotle. Therefore, I would appreciate it if you could explain this claim of yours, so I can understand what you are talking about. — Metaphysician Undercover

The modern understanding of causation as used by the sciences, might involve inductive reasoning, but isn't reducible to inductive reasoning. For example , if all ravens are black, then it must be the case that a sampled raven is black, but one wouldn't want to say that all ravens being black was the "cause" of a raven to be black. So induced hypotheses aren't causes per-se.

Inductive arguments are relevant to causation when one is reasoning about the type of an observed object when estimating how the object will behave , e.g when estimating whether an observed white ball is a snooker ball. But when deciding whether a particular relation between two particular events is a causal relation, induction cannot be applied if there isn't a general case to appeal to, yet the existence of a general case isn't said to be necessary for a particular causal relation to exist. So induced premises aren't necessarily causes, and causes aren't necessarily inducible.

Nowadays, an instance of a 'causal' relation between a particular cause A and a particular cause B, is understood to be relation which asserts that B occurs if and only if A occurs, assuming that nothing else could be the cause of B. This is what is meant by saying that causation involves "counterfactuals".

A scientist obviously cannot go back in time to test the truth of an alleged instance of a causal relation. Instead, he simulates his definition of the causal relation using model to see how simulated instances of that relation behave in comparison to simulated instances that aren't of that relation, and can at most conclude that if the alleged instance is of that relation, then the instance behaves in the same way as the other simulated instances. If the scientist always presents his conclusions as being conditionally true given the truth of the assumed hypotheses (induced or otherwise), then he avoids committing the fallacies of induction that routinely occur in the scientific literature.

Yes, physicists are actually heavily invested in the use of "causation". Take a look at the concepts of "lightcone", "timelike & spacelike", "worldline", "propertime", for example. They use knowledge of the temporal order of events (causation) to establish timelines in relativity based observations. — Metaphysician Undercover

As Russell observed, a temporal order per-se does not imply causation. For example, the orbits of the planets are describable by a differential equation that makes no appeal to cause and effect. The space-time manifold of General relativity makes no use of causation, nor does the evolution of a phase-space describing a dynamical system. More generally, a theory that sticks to describing actual phenomena, makes no mention of causality.

The concept of causation is actually a metaphysical interpretation of counterfactual logic, as extensively used in the design of double blind experiments. By definition, counterfactual outcomes aren't observed in experiments, so an interpretation of counterfactual logic that rests upon a speculated existence of non-realized experimental outcomes, cannot be verified through scientific experiments. But of course the social sciences do use counterfactual logic since they interpret the logic empirically, implying that the use-meaning of counterfactuals is in conflict with the traditional philosophical understanding of counterfactuals as literally referring to other possible worlds.


One of the ironies of the super-determinism interpretation of QM, is that it implies the non-existence of causality, since if reality is fully determined such that there are no counterfactual outcomes, then the resulting super-determined reality is merely a true story whose course of events is absurd.

According to the 17th Century Catholic priest Nicolas Malebranche, all effects are spontaneous whether or not they are attributed to causes, due to the fact that "Created things are at best "occasions" for divine activity. Bodies and minds act neither on themselves nor on each other; God alone brings about all the phenomena of nature and the mind" - Wiki

Malebranche can be interpreted as preempting Hume's conclusion that causal conditionals are not analytic, due to the fact that the effect of a causal relation isn't logically necessitated by the cause of the relation. Thus the effect of every causal relation must be a spontaneous act of god.

So from the point of view of the occasionalists, sponteneity is the essence of causality.

There is no a priori reason as to why the past should be either finite or infinite, for the past might be potentially infinite and grow in response to present and future observations. For there isn't a means of determining that past exists prior to, and independently of, the discovery of historical evidence.

No, that actually proves a first cause. "What caused a circular causation to exist instead of another type of causation?" As you noted it "Has no initial-cause", thus there is no prior explanation for its existence. Meaning, its a first cause as defined in the OP. — Philosophim

One can interpret circular causality as saying that there is no initial cause, or as saying that what is considered "initial" is subjective or relative to the observer. The important thing, is that causal circularity implies that every causal relation is symmetric and of the form A <--> B. or equivalently, that the causal order A --> B --> C comes equipped with a dual order in the opposite direction, C --> B --> A.

However, circularity isn't a requirement for symmetric causal relations. E.g the interpretation of QM known as "Super-determinism" is in effect committed to symmetric causal relations as a consequence of denying the existence of counterfactual measurements, without committing to temporal circularity.

Also, a presentist might interpret the present as being the perpetual "first" cause , in spite of also admitting that present events are caused by "past" events when speaking in the vulgar. To resolve this apparent contradiction requires distinguishing causality from temporality, including the topologies in each case that might conceivably be different.

Fixed-point iteration, i.e. F(z) = z, is the mathematical description of circular causation, which can be considered a non-finite conception of causality that is symmetrical and has no initial-cause, thus also eliminating the causal arrow.

That brains create consciousness? We've figured that out. — Philosophim

Did we figure it out in the sense of figuring out the truth of a proposition, or did we merely define "consciousness" as referring to what brains do?

It sounds like you are merely enumerating trivial tautologies that convey no information in being true-by-definition. Give us an example of a non-trivial analytic truth that qualifies as "knowledge".

Secondly, how can non-recursive analytic truth be said to exist? The purpose of recursive grammar is to put into place authoritarian rail-roads called "unbounded quantifiers" in order to show pretend dictate that new analytic 'truths' are derivable from old ones. If our analytic truths contains first-order arithmetic then we run into undecidability but at least have extendable rail roads, else analytic truth is reducible to quantifier-free decidable propositions that have no inferential or normative implications, such as a law of addition being defined but only for the first fifty numbers.

Isn't knowledge supposed to be informative or at very least serve as a normative fiction?

(Going further, how can knowledge be informative if it isn't fallible? Isn't the very concept of knowledge broken?)

The meaning of modalities is in their use, which is inadequately represented by picture-theories of modalities, especially the silly Venn diagrams stemming from the naive depiction physical possibilities as being a proper subset of metaphysical or logical possibilities.

First of all, are modalities empirical claims about reality, or they normative rules of convention that refer to the use and interpretation of a model, or are they both? And besides, how does the empirical content of a model relate to the application of it's rules? Can Kripkean semantics, or any other plum-pudding depiction of possible worlds do justice to the complicated use meaning of modalities?

Consider the fact that physical impossibility cannot be empirically falsified, at least not in the naive way that people presume. For example, the physical impossibility of faster than light travel cannot be directly tested nor understood by measuring the speeds of various objects, for we cannot observe what isn't observable, and the literal claim that nothing can travel faster than the speed of light cannot be directly verified by any finite number of experiments. Nor can a philosopher directly imagine faster than light travel in a thought experiment (for what would that look like, exactly?). So both the empirical and theoretical meaning of the impossibility of faster than light travel is far from straightforward and definitely not obvious. Furthermore, the literal English meaning of "faster than light travel" cannot even be translated into the language of Special Relativity, for SR maps the English sentence "faster than light travel" to infinite Lorentz factors that are extensionally meaningless.

The empirical meaning of SR is demonstrated by the experiment and results of the Michelson Morley experiment that partly motivated it. This empirical meaning does not refer in any obvious way to the sentiment that "faster-than light travel is impossible". If a physicist is asked to describe the meaning of this impossibility, he will likely refer to empirically observable Lorentzian relations that he argues are expected to hold between observable events. In other words, his use-meaning of the "physically impossible" is in terms of the physically possible!

So physical impossibilities shouldn't be thought of in terms of impossible worlds, but rather as referring to the application of a linguistic-convention that supports the empirical interpretation of language.

A language consists of a trade-off of semantic ideals that includes (among others) universality versus domain-specific authenticity, expressiveness versus efficiency, communicability versus idiosyncratic privacy, reliability vs adaptability.

I really don't understand what you are saying here. You appear to be saying that you see no clear distinction between past and future, because you interpret everything "within the context of the present". — Metaphysician Undercover

Yes, roughly speaking.

But isn't it the case that your reference to "the present" already implies a clear distinction between past and future? What could you possible mean by "the present", other than an assumed separation between memories of past, and anticipations of the future? Therefore your reference to "the present" seems to already imply a clear distinction between past and future. — Metaphysician Undercover

I understand the tenses to be closely related to modal distinctions made in relation to the present, but I don't deny the modal distinctions, nor the practical psychological distinction between past and future, or what McTaggart crudely referred to as the A series (is psychological time really a series?). But like McTaggart, I don't think the information content of the "A series" has any obvious relationship to the B series which is all that the public theory of physics refers to, or to the broader physical conception of time that Wittgenstein occasionally referred to as "information time" which i think of as a "use-meaning" generalisation of McTaggarts B series that also includes the practice of time keeping ( see Hintikka for more discussion on Wittgenstein's evolving views on the subject).

Furthermore, you refer to "present observations", but this concept is logically flawed. There can be no such thing as present observations because "to observe" is to take note of what happens, and this implies that an observation, being what has been noticed is necessarily in the past. It is this idea, of "present observations" which is actually self-contradicting. — Metaphysician Undercover

The word "present" is only used to stress the distinction between the A and B series and the fact that observations are always in the present tense, even when they are used to evaluate past-contigent propositions (which are understood to be past-contigent in the sense of the B series, but not necessarily in the sense of the A series)

So yes, observations are not of the present but they are always in relation to the present tense. Furthermore, if the B series isn't reducible to facts that are obtainable in the present-tense then the existence and usefulness of the B series can be doubted or denied, and at the very least cannot be reconciled with the the present-tensed practice of physics.

Put predictions aside for a moment. How would you deal with possibilities in the sense of "it is possible for me to do X, and possible for me to do Y", when X and Y are mutually exclusive? If I act for Y, then X is made to be impossible, and if I act for X, then Y is made to be impossible. However, at the time when I am deciding, both are possible.

How can we model this type of future in relation to this type of past, when both X and Y change from being equally possible in the future, to being one necessary, and one impossible in the past? What happens at "the present" to change the ontological status of these events? — Metaphysician Undercover

If I speculate that the past might change, then aren't I contradicting the very definition of what i mean by "the past"?

And If i speculate that the future is already decided, then aren't I contradicting the very definition of what i mean by "the future"?

I don't conceive of a clear distinction between the tenses and the modalities. I interpret both empirically within the context of the present, even I don't consider their meanings to be empirically exhausted by present observations, memories, intentions, actions and so on.

It doesn’t seem an apt analogy to me. At issue is the nature of the object in question and what it is that transforms it from a possibility to an actuality. — Wayfarer

Does it even make sense to consider the modalities (or tenses) to be the subject-matter of physics? For aren't the modalities the very essence of what is meant by an 'explanation' that are inevitably invoked when explaining any explicandum in any subject?

Unless physics is willing to collapse the explanans/explanandum distinction by appealing to circular reasoning (which for many would defeat the purpose of an explanation), then i cannot see how the metaphysical concepts of modalities can be treated as first-order physical propositions that warrant physical explanation.

From an instrumentalist perspective, scientific theories are conditional propositions that do not say how things are in themselves, but rather predict or describe the empirical consequences of performing a particular action or observation in a particular context. So according to this perspective, possibilities are what is directly expressed by scientific theories, but not what is represented or referred to by such theories.

That is true, but the nature of the object who's existence is only possible is not. And that is the point at issue in this context, as the putative object, a component of the atom, is supposed to be amongst the building blocks of material existence. — Wayfarer

If a weather-forecaster states that tomorrows weather is possibly heavy showers, i interpret his sentence to be an empirical report regarding his model of the weather, and not literally to be a reference to tomorrows unobserved weather. (In general, I don't consider predictions to be future-referring in a literal sense, for the very reason that it leads to conflating modalities with theory-content and facts)

Modalities only arise in conversation when a theory is used to make predictions. But the content of theories never mention or appeal to modalities, e.g neither the Bloch sphere describing the state-space of a qubit, nor the Born rule describing a weighted set of alternative experimental outcomes appeal to the existence of modalities. Rather the converse is true. E.g a set of alternative outcomes stated in a theory might be given possible world semantics, but the semantics isn't the empirical content of the theory and so does not ground the theory, in my empiricist opinion.

Possibility is an empirical notion. In the case of QM, possibilities either refer to directly observable interference patterns, or they refer to statistical summaries of repeated trials. It is also a good idea not to conflate the empirical meaning of possibility with the epistemic notion referring to possible world semantics, which refers to how people use and think about theories.

IMO, reifying possibility to the status of multiple actual worlds is a mistake born out of equivocating the various uses of the term.

A description is not the thing described. — Wayfarer

True, but the distinction is easily lost in communication.

To see a robot as a mind is not to infer that the robot has a mind. By contrast, to see that the robot has sensors relaying information to Machine Learning algorithms is not to see the robot as having sensors and ML algorithms.

The word "other" in "other-minds" is where the confusion lies, for insinuating indirect-realism with respect to the mental qualities that we directly project onto others.

Oddly enough, I believe it's correct. — Wayfarer

But presumably human cognition, emotion, awareness, and behavior are equally describable in terms of adaptive algorithms , data, environmental feed-back and pattern-matching.

In which case, how can disagreements over the sentience of chatbots, robots , non-human animals, and even disagreements regarding the sentience of other human beings, be regarded as disagreements over matters-of-fact?

Isn't the concept of other-minds reducible to the concept of empathy? In which case, the sentence " a rock doesn't have consciousness" isn't a proposition about the rock. Instead, it has the same meaning as "I cannot relate to a rock", implying that if the rock ever began to act like a human, then I would change my mind about the rock , and that my new opinion about the rock would not be in contradiction with my old opinion or with other people's contrary opinions.

(If the public disagrees as to whether a chatbot is conscious, are they really disagreeing over facts about the chatbot?

There is more to an ideal of reasoning than the ability to apply logic in a valid way. There is also the pattern recognition applied to diverse empirical observations that allow for recognition of false premises. For example the "training set" which is hugely important to the results yielded by modern AI. — wonderer1

Yes, very much so. The successes of Machine Learning generalisation are entirely the consequence of ML models evolving over time so as to fit the facts being modeled, as opposed to the generalisation performance of ML being the consequence of a priori and constructive mathematical reasoning, as if purely mathematical reasoning could predict in advance the unknown facts being modeled.

And yet many popular textbooks on ML written around the turn of the millennium presented the subject as if successful generalisation performance could be mathematically justified in advance on the basis of a priori philosophical principles such as Occam's Razor, Non-informative prior selection, Maximum Entropy and so on. Notably those books only very briefly mentioned, if at all, Wolpert's No-Free lunch theorems that put paid to the idea of ML being a theory of induction.

Anselms's ontological argument is mine, in spite of it's theological pretenses, for it is an example of a logically valid constructive argument that is 'necessarily true' but nevertheless draws a false conclusion about the world outside of logic, in spite of the argument insisting that it is referring to the outside world!

As I see it, the argument is but one of infinitely many examples of a logically valid but false arguments, that presents negative evidence with regards to the epistemological utility of constructive logic, and thus in turn presenting negative evidence regarding the epistemological utility of a priori philosophical arguments, such as transcendental arguments. In other words, even ideal reasoners can be expected to draw rationally "correct" yet empirically false conclusions about the world. In which case, what is the point of AI and cognitive science?

Does Permanence/Impermanence of the soul necessarily refer to a fact about souls, or might it refer to the grammar of the word "soul"? (Theology as grammar)

For example, consider a presentist who considers the concept of change to only refer to objects but not to subjects (since he believes the present to be the only moment of time). Then he might assent to the sentence that "the soul is permanent", as a vulgar way of expressing his view that the word "impermanent" isn't applicable to subjects.

But what if the object of translation was not optical redness but brain states? It seems then that the context problem doesn't apply because Mary's perceptions are always present alongside her brainstates and correlate so much that many suspect that they are identical. — Apustimelogist

I'm not sure what neuropsychology means by 'brainstates' exactly - but then isn't that the point - that the types and tokens referred to by neuropsychology are sufficiently vague and flexible so as to both accommodate the ad-hoc and informal judgements of it's practitioners on a case-by-case basis, whilst conveying enough of the practically essential information?

I'm also reminded of software-engineering, where the concepts of types, tokens and type/token identity are normative notions that only concern and describe the programming language being used, rather than being descriptive of the implemented application (that could be implemented in any number of languages that use different and incompatible type-systems).

In my view, Physicalism takes types, tokens and identity relations too seriously, due to mistaking these normative linguistic concepts for propositions.

Physics could dissolve any particular "hard problem" of consciousness, by simply expanding the rules of it's language to accommodate any perception, in a bespoke, albeit practically unworkable fashion.

For example, take the colour scientist Mary from the knowledge argument, who "learns" about redness for the first time when leaving her black and white room. Suppose that upon leaving her black-and-white room and seeing red for the first time, the language of physics is augmented with a new term that specifically denotes Mary's red perceptual judgements. Call this new term maryred. There is one simple rule for this new term ; whenever Mary perceives an object to be "red" then by definition the object is said to be maryred. So if another scientist is performing an optical experiment, say on a distant planet, and wants to know whether the result is maryred or not, then according to the definition of maryredness, there is nothing he can do other than to ask Mary after she has inspected the result for herself.

Mary cannot explain the relation between optical redness and maryredness, and the augmented physical language doesn't specify theoretical rules for inter-translating the two, not even when additional context is provided. But why should this absence of translation rules be considered a problem for physics? Isn't it in fact a blessing that we might call "The Hard Feature of Physics"?

For suppose that maryredness was theoretically correlated to optical redness (plus context). Then doesn't this imply that Mary needs to be present at every optical experiment performed anywhere in the world, including the ordinary optical experiments that aren't measuring maryredness? For how can it be argued that maryredness is theoretically reducible to optical redness + context, but not vice-versa? Theoretical translation must surely work in both directions. So wouldn't the meaning of optical redness become contingent upon the meaning of maryredness such that Mary's perceptual judgements became part of the theoretical foundation of optics? Clearly this isn't desirable, because we want physics to be a universally applicable language with a semantics that is independent of the perceptual judgements of particular observers. So it makes good sense for physics to decree optical redness and maryredness to be incommensurable by fiat.

Hence in my opinion, those who believe in a "Hard Problem of Consciousness" misunderstand the purpose of science, and that this hard problem is better understood as being a "Hard Feature of applicable Physics"

Recall that Euler's postulates weren't given in relation to a system of numbers; he took lines and points to be primitive concepts. Relative to his informal axiomatisation, the length of a hypotenuse is "real" in the sense that it is a constructible number, meaning that it can be drawn using the practical method of 'straightedge and compass', which is algebraically expressible in terms of a finite number of mathematical field operations.

When it is disputed that a hypotenuse has a "real length", it is when geometric postulates are used to interpret Euclidean space in relation to a fixed Vector-space basis. The irrational points of a Euclidean space aren't extensionally interpretable unless the basis of the underlying vector-space is rotated so as to transform those irrational points to rational values, which also leads to previously rational-valued points to become irrational. So the problem of incommensurability is really about the fact that it isn't possible to represent all points finitely at the same time, which implies that Euclidean Space cannot serve as a constructive logical foundation for geometry.

The obvious alternative is to follow Alfred North Whitehead in 1919-1920, and abandon classical Euclidean topology for a 'point-free topology' that refers only to extensionally interpretable "blobs", namely open-sets that have a definite non-zero volume, whose intersections approximate pointedness . Then it might be possible to extensionally interpret all such "blobs" in relation to a fixed basis of topological description in a more constructive fashion, meaning that extensional ambiguity is handled directly on the logical level of syntax, as opposed to on the semantic level of theory interpretation.

I think there are definitely problems with the main ways of defining probability, particularly frequentism, but I don't think circularity is one of them. https://plato.stanford.edu/entries/probability-interpret/ . — Count Timothy von Icarus

Probability Theory actually supports what i'm saying.

First recall that Classical Probability Theory is said to speak of 'events' of Probability 1 that occur almost surely, and conversely of 'events' of Probability 0 that occur almost never. So although classical probability is sound in the sense of comprising an identifiable class of entities belonging to the universe of, say, ZFC Set Theory, it's semantics is in contradiction with naive intuitions about chance.

E.g when probability theory is interpreted as saying that a dart must land somewhere on an infinitely divisible dart-board, at a location that has probability 0. One the one hand, we want Pr(1) to mean surely, and Pr(0) to mean never, but this 'exacting' demand conflicts with our other demand that it is possible to choose any member of an infinite set. What probability theory is actually expressing, is that our intuitions about chance, determinism and infinity are vague and contradictory and cannot be reconciled, let alone be formally represented in terms of a finite axiomatic definition.

An obvious way out of the above impasse is to interpret almost surely and almost never as referring to limits of a sequence of random events, such as the dart's sequence of positions over time, where these limits aren't considered to represents probability-apt events in themselves. In which case, we restrict our interpretation of Probability Theory as only assigning meaningful probabilities to either incomplete trajectories of darts that haven''t yet landed and whose eventual position is uncertain, or to landed darts whose position is vague and to within finite precision among a set of positions whose probability is strictly greater than zero. In my view, this way out amounts to a philosophical rejection of an absolute distinction between determinism and chance.

That's an interesting idea. Any tips on a place to read more? — Count Timothy von Icarus

Sadly I can't think of specific references off the top of my head, but in my view Category Theory is the right meta-language for relating physics, logic and philosophy, so Samuel Abramsky and Jean Yves Girard would be my generally recommended authors, Plus lots of nlab and SEP, of course.

It basically comes down to this; "If something is not determined by anything in what way is it not random? — Count Timothy von Icarus

I think that alternative interpretations of 'chance' is the key to non-classical compatibilism, where by "non-classical" I am referring to considerations from modern logic.

Consider the fact that the definition of chance appears to be circular - ordinarily, chance is taken to mean "to not be determined", where to be "determined" is taken to mean "to not be subject to chance".

One way out of this circularity is to consider determinism and chance to be relative to perspective, by taking inspiration from game-theory in which "chance nodes" are understood to refer to states of a game in which it isn't the player's turn to move, but someone else's.

Non-classical compatibilism that is based on this logic, can take metaphysical "free choice" as an axiom that is true for every player of the game, whose actions impose constraints on both the possible futures and possible 'pasts' of every other player. This position can be regarded as "compatibilist" to the extent that it can successfully reduce the empirical observations of modern theoretical physics in terms of a set of laws, whose 'determinism' is considered to be relative to the frame of reference used.

Transactional QM seems to be the closest theory in this regard.

A person can recognize that we are physically determined systems, and recognize that we are systems that develop probabilistic anticipations of future events. Furthermore, it's rather pragmatically valuable for machines like us to discuss such anticipations. (To get a job, to get married, to get to the moon, to end global warming, etc.)

It seems to me there is a pragmatic value, for the sort of machines we are, to being able to communicate in simplistic terms of free will, and as we are able, modify what we mean by "free will" to be more accurate. — wonderer1

Yes, but if determinism is accepted by the compatibilist, then probabilities can only be given an epistemic interpretation, while teleological concepts such as "anticipating the future" can only be objectively interpreted as referring to present and past causes. In which case, your pragmatic compatibilist solution must surely collapse on further inspection into standard metaphysical determinism without "free will".

Another possibility which comes to mind, is to deny that there is an absolute metaphysical distinction between determinism and free-will, by arguing that a definition of either is meaningless, by virtue of their definitions being circular. This is analgous to the arguments that Quine used to reject the analytic-synthetic distinction. However, since this is about denying the intelligibility of the determinism/free-will distinction, I can't see how this stance could be described as a "compatibilist" position. Furthermore, it entails re-conceiving the problem of free will as being at least partly grammatical in nature, as opposed to referring to a purely physical conjecture.

You said it like the compatibilist model of the world has retro causality, but I think instead it's more accurate to say that your model of compatibilism has retro causality. — flannel jesus

Retro-causality is a generally vague and controversial concept, to the point that it seems to rule very little in or out (recalling the fact that QM, which most physicists consider to be forwards-directed, has an innocuous retro-causal interpretation). Causal conventionalists like Hume for instance, even rule out retro-causality as a matter of tautology, which is why i didn't want to appeal to retro-causality as a hypothesis (which some might argue is formally meaningless), but to philosophical and empirical intuitions, naive if you like, that align with the idea.

It might have been better if I had never used the term. What is of underlying importance to compatibilism in my view, isn't the existence of retro-causation (whatever it is supposed to mean), but the treatment of material implication as being symmetric, i.e. of the form A <--> B, which can be interpreted in a number of ways, including Bertrand Russell's directionless "no causality" view, super-determinism and circular causality. In these cases, it is accepted that there exists synchronisation between a so-called "cause" and a so-called "effect", but where the control between "cause" and "effect" is either considered to be bidirectional, directional but a matter of perspective, or directionless in both directions.

I don't know the background motivation of the OP, but the problem that was presented is very reminiscent of the thought experiments that physicists use when selecting among interpretations of QM, which frequently give rise to debates over free-will in magazines such as the scientific american. In fact the OP's thought experiment is more or less identical to premises called "quantum conspiracies" , namely the premise that nature has already decided on the properties that physicists will measure, such that physics experiments cannot reveal anything about nature's properties.

Why? Says who? — flannel jesus

Yours truly. Tell me how i've gone wrong.

I still have no clue why you think compatibilism and retro causality have anything to do with each other — flannel jesus

Because according to classical understanding of causality, the past is both fixed and exactly determines the future, which prevents the possibility of free choice of any agent who lives above the initial cause.

Compatibilism doesn't make sense as a concept unless the past is in some way considered to be ontologically dependent upon the future. Being committed to the appearance of retrocausation isn't to be committed to retro-causation, and super-determinism might even be considered as appearing retro-causal.