In Marxist writing about dialectics, one often finds very general assertions. "The various seemingly separate elements of which the world is composed are in fact related to one another." "The whole world, natural, historical, intellectual, is... a process, i.e., as in constant motion, change, transformation, development." "All reality is constantly changing." "As soon as we consider things in their motion, their change, their life, their reciprocal influence on one another... we immediately become involved in contradictions." "Real change must result from any contradictory system."
If we take statements like these literally, the implications are very strong.
Take the claim that "all reality is constantly changing." Does that mean that the value of pi changes? Perhaps we should exclude it from the proposition, since we cannot really even conceive of what it would mean for pi to change. (Or at least I can't; maybe some mathematician has.)
But what about a more material number, the fine structure constant - which is related, among other things, to the speed of light? Physicists know more or less what it would mean for it to change, but from what I understand, most believe it doesn't. Or take protons. Does dialectics imply that they decay, rather than existing indefinitely? Some physical theories predict proton decay, but there is no experimental evidence for it.
Lastly, what about the laws of physics themselves? Rees quotes the biologists Levin and Lewontin to the effect that one of dialectics' key insights is that "the laws of transformation themselves change" - for example, the laws which govern economic dynamics under capitalism will themselves be different in another historical epoch. Does dialectics then imply that, say, if the equations of general relativity are valid today, they were not yesterday and will not be tomorrow?
I am not taking dialectics anywhere new by asking these questions. Engels' writing on dialectics is full of examples from biology, chemistry, and math. But his examples appear mostly intended to illustrate, and are picked to be easy. If you look instead for hard cases like these, where change, contradiction, etc., are at the very least non-obvious, dilemmas begin to arise.
I can see four ways of approaching the question of what dialectics might say about proton decay.
The first approach is the most straightforward - to embrace the scientific implications. Yes, one might say, protons must somehow transform themselves under the impulse of their own internal contradictions; even if dialectics cannot by itself give us a complete physical theory, it can give us certain assertions, like the inevitability of change, as a starting point.
I find this implausible on its face; it would imply a much closer relationship between physics and philosophy than we usually see in real life. One can list plenty more physical examples which at least seem to contradict dialectics in this interpretation - and in the best case, the theory could be no more certain than its contrarian implications for physics. Experiments which extend what we know about the minimum lifespan of the proton would, in challenging one premise of dialectics, thereby challenge even dialectical arguments about capitalism.
A second approach would be to re-describe the propositions of dialectics as tendencies rather than absolute laws. So, one could say, while we should expect everything to change, including elementary particles, we must allow for exceptions when evidence or reason requires them.
This is not very satisfying. It makes dialectics something much less useful than Engels' "science", Rees' "algebra", or Novack's "logic". It is basically impossible to disprove, since there is no way of quantifying the proportion of exceptions across physics, biology, philosophy, and history. And, it would make it strange to even speak of a coherent subject called "dialectics". If dialectics is a mere set of rules-of-thumb - "don't expect things to be static", "look for connections between parts and larger wholes", "look for internal causes of change" - what makes these tendencies special? There doesn't seem to be anything to distinguish these, as rules-of-thumb, from others, like one I once heard from a professor: "If a statement is true for the first three random, non-trival examples you check, it's probably always true."
A third approach to the problem would be to say that in expecting protons to decay, one would be drawing too specific an implication, and looking for the wrong kind of change. Dialectics tells us - one might argue - that everything in the world is in some way involved in a process of change, without telling us that any individual part of the world must change in any particular way. So - the argument might go - protons do change: they move in space and time, they gain and lose energy, they join and leave nuclei. Thus they accord with the predictions of dialectics. Why ask more?
The problem here is subtler, but equally fatal for the original claim of universality. It is that we no longer know exactly what dialectics does imply or even suggest about any particular question; if things must change only in some aspect or in relation to some given system, this does not tell us anything much in practice. And it is not merely a question of needing to use a materialist method, to always apply dialectics to concrete reality. If we say that dialectics only applies to some aspect of any given entity, situation, process, or structure, and we cannot in advance tell which, that is functionally equivalent to admitting that dialectics only applies sometimes, and we cannot in advance tell when.
The third approach thus collapses into the fourth and last approach, which is to admit that dialectics does not give us laws which really apply to everything, even as tendencies, and in particular that it has nothing to tell us about proton decay.
Can the collapse be avoided? I can think of one more distinction which an advocate of the third approach might try to make. That is to say that, while not every true way of describing something must be dialectical, the dialectical way of approaching something will always be more fruitful, getting at a dynamic which is somehow more crucial. So, while dialectics might not tell us whether or not a proton will ever decay, that inapplicability itself tells us that understanding proton decay is not the best way to begin understanding particle physics. Decay is a particular prediction, not the heart of the theory.
That argument may be plausible for the example we have been using, but it does not hold up in general. Take another case: gravitational orbit, e.g. the orbit of the Earth around the Sun.
One way to describe an orbit is to say that it is the result of two balanced opposing forces. Gravity pulls the Earth inward, while centrifugal force pulls the Earth outwards, so in the end it travels in an ellipse. This seems like a classical dialectical triad - two contradictory forces impelling motion in a new, third direction. But in fact, centrifugal force isn't real in the same way that gravity is, even in Newtonian physics. The Earth is not pulled outwards; rather, absent any external force, it would travel in an inertial straight line, and gravity merely bends its path. Inertia is not really a force. Moreover, according to the theory of relativity, even gravity is not really a force. Instead, space-time itself is bent. So the first "dialectical" view is accurate on the surface, but misleading from the perspective of a deeper theory.
I have focused so far on just one universal claim that dialectics might make, the claim that everything changes. There are other claims available to the theory. But without discussing them individually, let's note what's already been established, if, like me, one finds the modest fourth approach the most plausible.
A defensible version of dialectics will be a theory not about everything in the universe, but at most about processes of change, where we have already established that they exist. This isn't necessarily that much of a retreat, and it gives us a handy, non-circular definition of a dialectical "totality": not merely any arbitrary collection of nouns, but a coherent process of motion, development, or transformation - in the maximalist interpretation, any such process. The key question then becomes the concept of "contradiction", and its relationship to change and causality.
That question is worth its own post. But already in this one, we have not just slightly narrowed the territory dialectics might claim. We have also noted a dilemma that any conception of dialectics will have to face. Does the dialectic have specific implications for scientific theory, in any field, or not? Either answer is problematic; on the one hand, we risk an idealistic if not Lysenkoist attempt to make science answer to a 19th-century philosophical theory, and on the other hand, we risk reducing dialectics to a vague and insubstantial mysticism. As we explore what might be valuable in the dialectic, we will have to chart a course between these twin rocks.
 Rees, Algebra of Revolution, p. 5
 Engels, Socialism: Utopian and Scientific, chapter 2. More specifically, Engels says that it is the "great merit" of Hegel that he "represents" the world in this way.
 Novack, Logic of Marxism, p. 66. See also Marx's slightly different formulation: "The dialectic... regards every historically developed form as being in a fluid state, in motion" (cited in Rees, p. 100-1).
 Engels, Anti-Duhring, chapter 10.
 Rees, p. 118.
 ibid., p. 78.
 Protons are, after all, internally differentiated totalities; they are made up of quarks.
 Applied to statements about the natural numbers in a class on discrete math.