She doesn't, but that doesn't change the fact that you have a sentence in a human language on one side and the state of affairs which makes the sentence true on the other. And so the question is still how the snow being white makes the sentence white, because a sentence a state of affairs, no matter what theory of truth one espouses.

So deflationary theorists still have to account for how we know that the snow is white. — Marchesk

I don't think you meant to say "makes the sentence white", but rather "makes the sentence true".

Indeed, the deflationary theorists also have to discharge this burden. But I think they do, by means of broadly Kantian accounts of (intuition dependent) conceptual abilities and theories of judgment. (See for instance McDowell's Mind and World, or Sebastian Rödl's Categories of the Temporal.)

From a pragmatic realist (i.e., pragmaticist) standpoint, subjunctive conditionals are true when the laws of nature that they express are real generals; i.e., they are operative regardless of what anyone thinks about them. Peirce famously demonstrated this during a lecture by holding up a stone and stating that everyone in the audience knew that if he were to let it go, it would fall to the ground; and this was true even if he never actually let go of the stone. Similarly, a quality is a real possibility; e.g., if one were to shine broad-spectrum light on a red object, it would predominantly reflect it at wavelengths between 620 and 750 nm. Again, this is true even if no one ever actually conducts such an experiment. — aletheist

This is very similar the the suggestion that I made though it appeals to a conception of laws that rests on a metaphysics of events and Humean causation while my one suggestion was made on the background of a metaphysics of substances and powers (and substance/agent causation).

What's the difference between a deflationary and a non-deflationary correspondence? — Michael

The deflationary truth theorists make claims that sound very much like the things correspondence theorists say, but they chip away some of the metaphysically tendentious interpretations of those claims. For instance, a deflationary theorist might happily acknowledge that "snow is white" is true (or can be used to express a true proposition, in English) if and only if snow is white. But she doesn't claim this to imply that there must exist two metaphysically distinct sorts of things -- abstract propositions on the one side, and concrete elements of reality (i.e. states of affair) on the other side -- that somehow problematically correspond to one another. So, I wasn't defending correspondence theories, but merely suggesting that a correspondance theorist might make use of a strategy similar to the one a deflationary theorist might make use of to explain in a realist fashion the meaning of counterfactual conditional statements.

Perhaps something like the coherence theory of truth is better-equipped to handle them. — Arkady

Maybe one the main problems afflicting the correspondence theory of truth is the way in which it seems to presuppose a form of uncritical metaphysical realism. If the "holding" of a state of affairs, and the truth of the proposition somehow expressing this state of affairs, merely give rise to some sort of a "correspondence" of the former with the latter, then the issue of the conceptual structure of "reality" that makes this correspondance possible is rendered problematic. On the other hand, some apparently innocuous statements of the so called correspondence theory could be construed on the lines of a deflationary (or 'identity') theory of truth. This is how some philosophers (e.g. Jennifer Hornsby or Sebasian Rödl) construe Aristotle's claim that “to say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, and of what is not that it is not, is true” (1011b25). Thus construed, the theory doesn't entail nor presuppose any sort of problematic dualism of (conceptualized) propositions and (unconceptualized) states of affairs.

Back to the original question, then, if we admit of conceptually structured states of affairs, what "corresponds" to the truth of counterfactual conditional statements could be their being logical consequences of the things having real albeit unactualized powers. If our metaphysics admits of objects that have among their real properties not only "occurrent" qualities (either "primary" or "secondary") such as geometrical shapes or color, but also real albeit unactualized powers, then the proposition that "if I were to strike this porcelain dish with a hammer, then it would shatter" could be said to be true if the ascription of a real power, of which it is a logical consequence, "corresponds" to a truth about this power (in the deflationary sense of "corresponds"). That is, to say that it is true that a dish is liable to shatter when struck -- ascribing some sort of a "passive power" to the dish -- just is to say that it is liable to shatter when struck. And if that is the case, then any proposition that logically follows from this also is true, and this includes a definite range of counterfactual conditional statements. It is then the truth of this whole range of counterfactual statements that "corresponds" (in the deflationary sense) to something's having real unactualized powers.

Wouldn't it be more accurate to say that SOME of the laws of chemistry are approximations? — tom

I did not suggest that they were approximations, though some undoubtedly are. For a law not to apply universally need not entail that there must exist a more precise, as of yet unknown, "universal" law that it approximates. (This incorrect expectation lays at the root of Davidson's nomological principle of causality, I think)

My point was different. There are strong arguments to be made (on Kantian/Aristotelian grounds) that any law that purportedly governs empirical phenomena either must have exceptions (i.e. can be interfered with by something (or may fails to apply at some energy scale, etc.) or isn't really an empirically significant law but rather merely is an idealized abstract principle (a mathematical constraint, for instance). But, in the latter case, such a principle must always be interpreted in a determinate empirical context in order to be rendered relevant to the regulation of empirical phenomena that show up within specific domains of inquiry. We are then back to the first situation.

Truly exceptionless "laws" always are unreal abstractions, on that view. This possibility may be obscured by the tendency to conceive of "universal laws of nature" against the background assumption of a metaphysics of temporally instantaneous (and ontologically self-contained) "events" and Humean causation. But empirical results gathered from real experimental setups (and from ordinary human perceptual experience) always must be made sense of on the quite different background of a metaphysics of objects (i.e. traditional "substances") and their specific fallible powers.

Bohr's complementarity principle, in its wider philosophical generalization beyond the narrow scope of quantum mechanics, constitutes an explanation of this, I think. The principles of quantum mechanics determine not only the unitary evolution of pure quantum "states" (this is the abstract "universal" part of the theory) but must also specify the projections of those states onto definite "observables". And the latter always must make at least tacit reference to definite macroscopic experimental setups, as well at to definite human conceptual understandings, and/or pragmatic uses, of those setups.

What does it mean for an explanation to be complete? We are talking about science are we not? — Frederick KOH

Maybe you and I are, but Weinberg isn't. He takes his observation about the apparent convergence of the arrows of explanation produced by science to furnish irresistible evidence for what he takes to be a fundamental fact about "the way the world is". This is why his "final theory" could remain just a "dream", in his view; and "grand reductionism" would still be a true characterization of the ways the world is. The complete explanation at issue amounts to "reduction in principle", not the actual production of complete reductive explanations. You and Rorty may be Deweyan pragmatists. Weinberg isn't.

What Weinberg argues for is effectively the complete determination (which is a matter of metaphysics rather than epistemology) of the principles governing higher-level phenomena (e.g. animal behavior) by the principles governing low-level phenomena (e.g. particle behavior). If this were the case, then, there would exist complete reductive explanations that are knowable in principle, even if they were only really known to God.

My point was that you can occasionally go sideways and still converge. — Frederick KOH

Fair enough. But you can also fail to converge. Weinberg believes it to be an empirical fact (regarding the history of science) that actual arrows of explanation produced by the various special sciences tend to converge in the general direction of particle physics. But he focuses almost exclusively on reductive explanations and assumes without argument that non-reductive explanations that aren't pointing "down" to some underlying theory must be discounted as resting on uninteresting historical accidents (e.g. the contingent fact that an animal species has evolved in this or that way). So, his argument for "grand reductionism" is circular and the product of a lack of imagination (and, presumably also, a lack of familiarity with the analysis of the structure of non-reductive explanations.)

So what is the term you would use when the question "why do elements have the valencies they do" is answered by a theory of quantum mechanics? — Frederick KOH

If the explanation were complete (which I may assume it to be, for the sake of the argument), then the term that I would use is "reductive explanation". Ernst Mayr, in his book What Makes Biology Unique, offers many historical examples of successful reductive explanations in biology. I don't know how many time I must point out that for pluralists/emergentists to deny that *all* genuine explanations of scientific laws are reductive (as Weinberg would require them all to be) doesn't entail that they don't believe any explanation to be reductive. Some high level laws represent cases of strong emergence, and some are reducible. Weinberg wants them all to be reducible, because the only alternative that he can imagine is magic.

In any case, to get back to your specific example of atomic valence, it may be doubted that the explanation of such properties is reductive in kind. See the enlightening discussion of orbital hybridization in the case of the tetrahedral geometry of methane, in section 8 of Michel Bitbol's Downward causation without foundations (There is a direct link to the pdf file down the page.)

Converge. — Frederick KOH

They must converge towards a single point, according to Weinberg. They must converge towards the "final theory", somewhere underneath both the Standard Model of particle physics and General Relativity.

But if it can straddle multiple autonomous laws, why not also admit the objects of the theories of physics? — Frederick KOH

Why not indeed! But the pluralist/emergentist admits readily of them. Remember Ernst Mayr on "analysis"? For one merely to be admitting of the existence of such "reductive" (i.e. analytical) explanations amounts to what Weinberg downplays as mere "petty reductionism". What Weinberg's "grand reductionism" requires is that only the objects of the theories of physics be admitted, or, at least, that only the objects of theories somehow "closer" to the "final theory" of physics be admitted.

So do you agree with this

Valid claims and questions can be made within chemistry that straddles multiple autonomous laws. — Frederick KOH

For sure. It's Weinberg who would object to this, not me. He can't admit of his "arrows of explanation" pointing "sideways" to autonomous laws rather than them all pointing "below" in the general direction of his "final theory"

What do you call the framework which provides the vocabulary to express it and also the conventions used to determine its validity? And is this framework autonomous? — Frederick KOH

I never said anything about the autonomy of "frameworks". I argued that some theories are non-reducible when they explain some emergent phenomena and when some of their explanatory successful laws are autonomous in the sense that they are derived from the joint realization of specific sets of high-level structural features of the systems -- whole equivalence classes of them, as previously explained -- in which those laws apply. (There is an intentional circularity, here). Those high level structural features thereby circumscribe the scope of applicability of the emergent laws.

So, chemistry may be reduced to Quantum mechanics (specifically the Standard Model) plus initial conditions of the universe, plus several arbitrary constants, plus General Relativity (in order to provide the conditions for atom formation), plus thermodynamics, at least. Not much of a reduction! — tom

That's true, for sure. But, in addition to this, many of the laws of chemistry are valid only for some specific classes of bounded chemical systems (and presuppose specific boundary conditions, such as the total energy of the system). In that regard, such laws are akin to the unified sets of laws (or norms) of animal physiology and behavior that only apply to specific animals.

But surely you recognize that the situation in chemistry is very different. There is no specific law of chemistry with the reach and scope of QED. — Frederick KOH

Yes, and so? QED may have a wide range of applications and a large domain of validity. This has nothing to do with the question whether or not it might be reducible to some other, more "fundamental", theory.

How do you even stay within a single law when talking about a non-trivial experiment.

How did I "stay within" a single law (whatever that means)? I'm providing counterexamples to Weinberg's imprudent generalization one at a time.

Then I am not sure how to use your terminology here. What are what you call "high level structures" then? Are they logically different for each specific law even it they refer to the same sort of objects? — Frederick KOH

I don't recall using the phrase "high level structures" (though that might aptly characterize the objects and properties directly being governed by the laws of a high-level theory). I once mentioned high-level structural features, which include such things as the boundary conditions of a system, and which can serve as a theoretical basis for the direct derivation of the emergent (irreducible) laws of a high-level theory. Those high level structural features thus constitute the specific conditions under which the phenomena being explained and governed by the high-level theory can arise. They are being pointed at by the multitudinous "arrows of explanation" that mess up Weinberg's "grand reductionism" of explanatory-arrow-convergence.

Do you consider chemistry autonomous from the theories in quantum mechanics? — Frederick KOH

This binary question is much too crude. There are specific laws of chemistry that are autonomous with respect to the laws that govern simple molecular interactions. The situation here is much more complex and disorderly than it is in the case of the relation between to neighboring effective field theories that merely differ with regard to the energy scales of their respective domains of application. I made mention of a specific example concerning networks of chemical reactions a couple of days ago, though I didn't dig up the reference. Refer to Earley's discussion of "concentration robustness" in chemical reaction networks that satisfy the conditions of the theorem proven by Shinar and Feinberg (Structural sources of robustness in biochemical reaction networks, Science (2010) Mar 12). That's in Earley, Joseph E. Three Concepts of Chemical Closure and their Epistemological Significance.

The main point is that autonomy of a scientific discipline with respect to another always is autonomy in some specific respects. Some laws of chemistry are emergent, some aren't.

So there is a directionality between the two, leaving aside what to conclude from this directionality. — Frederick KOH

Yes, there is.

Example "why is the photon massless" is question expressible in terms of QED. — Frederick KOH

Yes, sure. The photon's being massless is a requirement for QED being renormalisable. (I didn't remember that, by the way. I Googled it). What's your point?

In every theory there are open problems describable in terms of the theory itself. Does this apply to what you call autonomous theories? — Frederick KOH

I can't say. Your first sentence is too vague. Maybe you could phrase it more precisely and explain the relevance of your question to what we've been discussing.

Let's get one thing straight first. While EWT is a theory for energies above 246 GeV, it is also for energies below that. In other words it is not illogical to say that it is an alternative theory to QED at energies below that.

Do you disagree? — Frederick KOH

It's rather misleading to call it an alternative to QED when QED can be logically derived from it. It would be better to say that it's a more determinate theory. It is inferentially stronger and hence more falsifiable. (It defines a wider "exluded zone", John Haugeland would say; some potential experimental results that it rules out aren't ruled out by QED).

By the way, I just found out that Karen Crowther recently published a book: Effective Spacetime: Understanding Emergence in Effective Field Theory and Quantum Gravity

Yes. Because what you say is a bit unexpected.

You define an equivalence class in terms of an existing theory.

You do not define it in terms of a set of empirical data to explain

Have I interpreted you correctly? — Frederick KOH

I'm not sure what the difficulty is. There are parameters of the electroweak theory (EWT, for short) that can only be determined empirically through performing experiments in large particle accelerators (through producing and analyzing interactions that involve energies above 246 GeV). If this were not the case, then EWT would not be underdetermined by the experimental evidence that supports QED, and that is available through observing common lower energy interactions outside of particle accelerators. So, if EWT is considered to be one possible "reduction base" of QED (i.e. one that is actual, and not merely possible), then other similar theories that would have alternate values to the empirically determined (at high energy) parameters of EWT would constitute other possible reduction bases of QED. In other words -- and this is the main point -- the actual values of those determinate parameters of EWT are irrelevant to the explanation of the structure of QED, or to the determination of its specific laws.

So this is the criteria for being in the same equivalence class as QED: — Frederick KOH

No, QED is not part of the relevant equivalence class. You are still badly misconstruing what I said. Are you really incapable of reading a whole sentence? Can't you at least try to make sense of it with reference to the surrounding context of the discussion? Here is the whole sentence, together with the sentence immediately before it:

(PN:) "The theory of the electoweak interaction (i.e. the effective quantum field theory that is found to be empirically valid, as well as theoretically adequate, above the 246 GeV unificaton energy) is underspecified by the theory of quantum electrodynamics. All the alternative theories that would have been consistent with the validity of QED at the lower energy scale (i.e., any energy below 246 GeV) belong to an equivalence class of theories, such that QED can be derived from any one of them."

The alternatives that are being considered are alternatives to the theory of the electroweak interaction; not alternatives to QED itself! It is a class of possible realization bases (analogous to material realization in classical physics or other "high-level" natural sciences) of QED that Karen Crowther meant to specify. If this may help, as an analogy: if you were to define as belonging to the same equivalence class all the hats that fit on your head, then you yourself wouldn't be a member of this equivalence class. You should probably go back and read again the whole post rather than pulling out sentence fragments out of it, because the inferences you are drawing from those fragments are nonsensical, and lead you to miss the point by a mile.

Why? Some of that data was in existence before QED was even close to being a mature theory. — Frederick KOH

The set of theories that I mentioned are singled out as being part of the relevant equivalence class. It was meant as a definition of this equivalence class. You quoted only the first part of the sentence where I explain this and then suggested: "Wouldn't it be more accurate to say all theories (QED included) would have been consistent with experimental results at the lower energy scale". But that's not more accurate. That's saying something else entirely, quite irrelevant to what I meant. I had assumed you meant that I had forgotten to include QED in the relevant equivalence class. It need not be included at all, although, trivially, of course QED also is consistent with the experimental results that it itself has been devised to explain. It looks like, through reading only one half of a sentence of mine, and rushing to respond to it, you misconstrued what it meant.

Since it is an empirical theory, what experimental data is it consistent with? — Frederick KOH

You can read Feyman's popular "QED" book, if you're curious; or Google the Wikipedia page, maybe.

Is there anything you would disagree with here? — Frederick KOH

No but it's not a paraphrase of what I wrote. It ignores the point about underdetermination.

Wouldn't it be more accurate to say all theories (QED included) would have been consistent with experimental results at the lower energy scale — Frederick KOH

No. That wouldn't make sense. QED is not part of its own set of higher-energy (and shorter-range structure) possible realization bases.

While underdetermination is well known enough in the philosophy of science, could you give a central text which uses the term underspecified. — Frederick KOH

I meant underdetermined, thank you.

All the more egregiously in the case of Sean Carroll, who is, after all, a physicist. — Wayfarer

It's also a bit disappointing on account of the fact that Carroll, unlike colleagues of his like Hawking, Krauss or Weinberg, isn't utterly dismissive of philosophy. In the comment section of the first one in a series of four blog posts about emergence written by Massimo Pigliucci, George Ellis takes Carroll to task on this very issue (i.e. he points out the inadequacy of Newtonian billiard ball models as a basis for an anti-emergentist argument). Ellis proceed to discuss superconductivity as an example of a emergent theory where the very nature of the "low level entities" (in this case, Cooper pairs) is conditioned by high-level features of a physical system.

As I had suggested, however, even if deterministic Newtonian physics had turned out to be true, strong emergence would still make sense. Werner Heisenberg indeed understood early on, prior even to the development of his matrix mechanics, that if point particles obeying the laws of Newtonian mechanics were somehow conceived to be the material basis of the physical world, it would still not be possible to know determinately all their intrinsic mechanical properties, and thus those properties would become 'noumenal', as it were, and not allowed to be represented as genuine physical magnitudes in an empirically grounded physical theory. (I think that was in The Physicist's Conception of Nature, which I had read in the French edition some 30 years ago, so my memory is fuzzy. But the gist of his argument stuck with me.)

I think this misses the point.

Explanations at any level of emergence can be fundamental. We think of quantum mechanics and general relativity as "fundamental", which they are, but NeoDarwinism and the Theory of Computation are also fundamental.

There is no downwards or upwards causation between fundamental theories. — tom

If for a theory to be fundamental means that it is universal and applies everywhere, at any time, and on every energy/spatial scale, then very few theories are fundamental (not even general relativity). If it means that they provide autonomous explanations that abstract away from features of the contingent material constitution of the entities that they regulate, then stating that they are fundamental doesn't entail anything more than stating that they are autonomous. Hence, I prefer the term "autonomous". Quantum mechanics is more of a framework than it is a theory. It consists in a set of formal features shared by more determinate empirical theories such as quantum electrodynamics. Such theories are likewise autonomous.

The ideas of downward or upward causation don't relate to causal links between the theories themselves which regulate phenomena. They rather pertain to causal relationships between phenomena that belong to distinct levels of description, or scales of intervention. When an intervention on a macro-scale variable reliably produces a specific, targeted, effect on the micro-physical state of a system, for instance, that constitutes an instance of downward causation. The existence of meaningful downward causation is being disputed by some philosophers (such as Jaegwon Kim) and some scientists (such as Sean Carroll) on the same grounds on which they also dispute the existence of strong emergence. They believe both the ideas of strong emergence (and hence of the autonomy of "high-level" theories with respect to "low-level" ones) and the existence of (irreducible) downward causation to be inconsistent with the causal closure of the micro-physical domain. Those objections at least make some sort of intuitive sense in the framework of deterministic classical mechanics, but they are invalid, in my view, in a way that is simply made even more salient by their failures to go through in the context of quantum physics.

So you consider all these autonomous high level theories. In the case of quantum electrodynamics, electroweak theory is not a reduction, "since those higher-level laws are completely insensitive to any other low level features of material constitution that aren't merely deducible from the system's belonging to the relevant equivalence class."

Or did you mean something else? — Frederick KOH

No, that's exactly what I meant. The theory of the electoweak interaction (i.e. the effective quantum field theory that is found to be empirically valid, as well as theoretically adequate, above the 246 GeV unificaton energy) is underspecified by the theory of quantum electrodynamics. All the alternative theories that would have been consistent with the validity of QED at the lower energy scale (i.e., any energy below 246 GeV) belong to an equivalence class of theories, such that QED can be derived from any one of them. But the empirical discovery that one specific theory happens to be empirically adequate above 246 GeV (and up to the grand unification energy of the true GUT theory, presumably), adds nothing to the explanation of the validity of the laws of QED over and above this "reducing" theory belonging to the aforementioned equivalence class. This is where the (partial) explanatory autonomy of QED comes from and why the "arrows of explanation" stemming from Weinberg's question as to "Why?" its laws obtain don't point to specific features of the empirically valid theory of the electroweak force one step up (in order of increasing energy scales) in the hierarchy of effective field theories.

Then are these autonomous high level theories empirical theories? — Frederick KOH

Yes. The ideal gal law is an empirical law, and so are quantum electrodynamics or quantum chromodynamics (both of the latter are effective field theories), for instance. Ethological accounts of animal behavior also are empirical. The number of examples from natural or social sciences is almost infinite. Theories that are fully reducible are the exception rather than the rule.

A further point of clarification. You to refer to "relevant equivalence class" because a single high level theory may have instantiations with different low level features/substrates. Or to use a previous example, the same software can run on different kinds of computers.

Or did you mean something else. — Frederick KOH

No, that's broadly correct. I get the "equivalence class" concept from George Ellis, mainly. (And I also homed in on it independently in a manuscript titled Autonomy, Consequences and Teleology that I wrote in 2009). It is closely connected to Karen Crowther's idea of "universality", which alongside "autonomy" characterizes emergent phenomena and the emergent norms and/or laws that govern them. It is also closely connected to the idea of "multiple realizability" made use of by functionalists in the philosophy of mind. When a functional system is multiply realizable, then all the possible realizations fall under an equivalence class defined by the functional specification.

But multiple realizability, thus conceived, just is one sort of case where Crowther's more general idea of "universality" is exemplified by emergent phenomena in nature; and hence just one sort of way to characterize equivalance classes of systems that admit of the same high-level explanation. The other two cases consist in (2) equivalence classes of micro-physical realizations of a macro-variable (you can refer back to my discussion of the ideal gas law earlier in this thread, for instance) and cases where the underlying physical theory is underdetermined by the emergent theory. This is what is exemplified by the cases of critical phase transitions or broken symmetries that characterize, among other things, the relations between adjacent "effective field theories" that have their domains of validity tied to distinct energy scales. The phenomena of superconductivity and superfluidity exemplify this.

So attempting to synthesize your position: while "those higher-level laws are completely insensitive to any other low level features of material constitution that aren't merely deducible from the system's belonging to the relevant equivalence class", we can seek to explain how low level features enable the high level ones and such enabling explanations are genuinely reductive. — Frederick KOH

Yes, for sure. But this merely amounts to material constitutive analysis; something that Ernst Mayr, for instance, readily acknowledges as an important area (albeit just a part) of fruitful scientific inquiry, and that Weinberg tends to downplay as mere "petty reductionism" as contrasting with his own metaphysical claim of "grand reductionism" to a unique "final theory". But Weinberg is also blind to the positive features of the emergent relations (involving 'autonomy' and 'universality', as explained by Karen Crowther) displayed alongside reductive analysis. Those explanatory relevant positive features of emergent phenomena just destroy Weinberg's grand metaphysical claim since they give rise to "explanatory arrows" that point away from his dreamed of final theory.

Does Weinberg give similar caveats for his version what a fundamental theory is? — Frederick KOH

Not so far as I can see. Also, my objections can't be met with mere caveats. What comes closest to caveats in Weinberg's two texts are his acknowledgement that high-level theories are useful irrespective of them actually having been effectively reduced. But his reductionist claims are explicitly metaphysical rather then methodological/pragmatist. So this weak caveat isn't really relevant to my objection to his stronger metaphysical claim.

Another qualification that he offers is rather more akin to an anti-caveat. He distinguishes explicitly his own brand of "grand reductionism" from the more ordinary "petty reductionism" that he identifies with Mayr's "analysis" of a system onto material constituents in order to highlight bottom-up principles and constraints. Pluralists and emergentists are happy to recognize the explanatory fecundity of such a process of analysis. Weinberg is insistent that his own brand of theoretical ("grand") reductionism is much stronger than that. But it is owing to the strength of his claim that the counterexamples to it exemplified by the clear cases of strong emergence, ubiquitous in normal scientific practice, are fatal to it.

— Wosret

"The consensus of the sages — I recognized this ever more clearly — proves least of all that they were right in what they agreed on: it shows rather that they themselves, these wisest men, shared some physiological attribute, and because of this adopted the same negative attitude to life — had to adopt it. Judgments, judgments of value about life, for it or against it, can in the end never be true: they have value only as symptoms, they are worthy of consideration only as symptoms; in themselves such judgments are meaningless. One must stretch out one's hands and attempt to grasp this amazing subtlety, that the value of life cannot be estimated. Not by the living, for they are an interested party, even a bone of contention, and not impartial judges; not by the dead, for a different reason. For a philosopher to object to putting a value on life is an objection others make against him, a question mark concerning his wisdom, an un-wisdom. Indeed? All these great wise men — they were not only decadents but not wise at all." -- Nietzche

In that case, that's quite an anti-climax. Engineers create structures like this all the time. Engineers who make parts and components at one level are also at the same time creating abstractions for engineers at the next level. — Frederick KOH

Both the hardware and software levels are abstracts levels. (They're akin to the levels of cell physiology and of whole organism physiology). They also both are real functional levels. The software-level description characterizes real material processes (executions of high-level source code programs) that are both symbolically significant and that abstract away from some features of hardware (or virtual machine) implementation. The hardware-level processes (executions of individual steps of machine code, or, at an even lower level, of elementary binary logical functions) have the same features but what this level abstracts away from are the high-level symbolic modes of operation that are conferred to it (as viewed from above) by the higher level algorithmic structure defined by the software "loaded" into the computer.

Engineers may create structures like this all the time, as you notice, but so does nature. In some cases, the levels may be clear cut, as are the different energy scales in effective field theories, or, in other cases, fail to form neat hierarchies, as is the case for the Earth climate system or for biological organisms. What is common to all of those natural and artifactual (and also cognitive and social) phenomena is the ubiquity of downward causation and their illustrating the inadequacy of Weinberg-style reductionism as a purported description of "the way the world is".

The part of your difference with Weinberg where he does not consider
this autonomy to be fundamental - well I am on his side on this,

So, contrary to what Weinberg believes, the level of his hypothesized "final theory" isn't fundamental as a matter of "the world being the way it is", where this is conceived as stemming form all the "arrows of explanation" being found empirically to converge towards one single theory of particle physics. It rather consists in this lowest "material" level being dignified by him with the word "fundamental" merely through downgrading as not being really "fundamental" (or as being dependent on mere "historical accidents") all the real arrows of scientific explanation (a majority of them, actually) that don't happen to point towards his favored theory. The real phenomenon of the (partial) autonomy of higher level processes and phenomena from their lower level bases of material constitution just destroys his mythical structure of arrow convergence. This grand structure, rather than being a reflection of anything empirically verifiable in nature, turn out to be a product of his prejudice.

It is not a difference in the understanding of the facts. It is one of perspective.
I am sure you can debate perspective, but I would rather debate something else.

So, you really are after an understanding of the facts that doesn't rest on any conceptually informed perspective at all? Or is there a way to do science without making use of any theoretical or empirical concepts?

But there is another error of reductionism, which maybe even deeper: the misconception that our theories form a hierarchy. — tom

Yes, I thing that is true also. Causal networks in complex dynamical systems can be very messy and fail to display clear cases of upward and downward causation operating between neatly distinguished levels. (That doesn't prevent such messy systems from displaying stable attractors such as the deterministic surface warming response to the enhanced greenhouse effect.) The same is true of biological systems. Alan C. Love has written a fascinating paper in which he criticizes the narrow focus of theoreticians on neat hierarchies: Hierarchy, causation and explanation: ubiquity, locality and pluralism. The abstract may be worth quoting in full:

"The ubiquity of top-down causal explanations within and across the sciences is prima facie evidence for the existence of top-down causation. Much debate has been focused on whether top-down causation is coherent or in conflict with reductionism. Less attention has been given to the question of whether these representations of hierarchical relations pick out a single, common hierarchy. A negative answer to this question undermines a commonplace view that the world is divided into stratified ‘levels’ of organization and suggests that attributions of causal responsibility in different hierarchical representations may not have a meaningful basis for comparison. Representations used in top-down and bottom-up explanations are primarily ‘local’ and tied to distinct domains of science, illustrated here by protein structure and folding. This locality suggests that no single metaphysical account of hierarchy for causal relations to obtain within emerges from the epistemology of scientific explanation. Instead, a pluralist perspective is recommended—many different kinds of top-down causation (explanation) can exist alongside many different kinds of bottom-up causation (explanation). Pluralism makes plausible why different senses of top-down causation can be coherent and not in conflict with reductionism, thereby illustrating a productive interface between philosophical analysis and scientific inquiry."

I suspect this autonomy is the autonomy that computer designs at the logical level have. It just happens that economics and technology has determined they be implemented using semiconductor technology. But the design does not depend on it.

Is this a correct paraphrase? — Frederick KOH

That would be a relevant example. We may say that the software laws govern how the computers behave, at the relevant functional level that gives meaning to significant input/output structures. The lower levels of hardware implementation enable rather than govern what the software does (as characterized at the relevant symbolic level). In the case where a bug can be traced to a hardware malfunction (e.g. a real winged bug being fried up on a vacuum tube) rather than to a programming error, and only in that case, are reductive explanations of the episode of software failure to meaningfully be sought after. We can also seek to explain how the software is being enabled to run effectively on a specific machine, and such enabling explanations are genuinely reductive. But they are answers to a different question, and not even indirectly relevant to the high-level question concerning the obtaining of the input/output structure that is fully explained by the software specification. (Indeed the hardware design might plausibly have been implemented by the computer builders in order to enable the execution of the software in accordance with its own autonomous laws. So the ensuing operation of the hardware, under the governance of the software instructions, constitutes a clear case of downward causation.)

Pointless is not impossible. — Frederick KOH

It is pointless because it is impossible. It is also pointless because, even if, per impossibile, such a reductive explanation were to be achieved, it would be redundant with the formal explanation at the emergent level. You would have a case of causal overdetermination or epiphenomenalism. More plausibly, you would have a case where the reductive "explanation" is an overly complicated pseudo-"explanation" that is entirely parasitic on the high-level explanation, since it would merely amount to claiming that the material constituents were determined by low level laws (and initial conditions) to come to realize an arrangement that happens to constitute the system's falling under the high-level concept that is a causal antecedent for the high-level law such as to thereby necessitate a micro-physical arrangement that (as it happens) realizes the consequent of the law. But that is just another way of stating what the high-level law already was understood to necessitate and fully explain in the much simpler high-level terms while abstracting away from the micro-physical features of the system that contribute nothing to the causal explanation.