Jump to content
Objectivism Online Forum

Rate this topic

Recommended Posts

I have a fairly simple problem / question / or need (let my need become a demand on your attention!): what is a theory? From experience, I know a number of specific theories, but I do not know what the proper definition of “theory” is, and what its properties are. My ultimate goal is to say something about a particular scientific theory (to identify flaws stemming from a misunderstanding of what a theory is). To show this, I need to say what the essence of a “theory” is. By analogy, I know what the concept “concept” is. Knowing the nature of a “concept”, I know that “1967 Dodges, black cats and the act of running” –excluding all other things – cannot be a concept, since those things have no similarity.

I confess that I have a draft of a theory of “theory”, in the more literal scientific or philosophical sense (thus excluding uses where someone says that they “have a theory that X”, when they mean that they “feel that X is so” or they “have an idea that X may be true”). A theory is (defined as) a system of identifications which allow man to grasp the nature of a (conceptualized) subject. It presumes a definition of the subject concept, thus “theory of gravity” presumes a concept “gravity”, which implies a definition of “gravity”. Likewise “theory of mammals” presumes a concept “mammal” (and therefore a definition of “mammal”). A theory of a subject is a set of (highly) probable propositions which state the essential properties of that subject. The underlined parts here are my theory of “theory”.

I need to clarify a few points. A “property” of a thing is a fact about its composition that determines what it does, which is not the same as “an observation” or “a correlation” true of the thing. For example, Android is the most popular OS for smartphones, but this is not a property of Android. Plutonium is used in reactors and making nuclear weapons, but this is not a property of plutonium. As for “essential”, I first want to disclaim any connection to discussions of essential vs. accidental properties in professional philosophy, which gets bogged down in proper names as opposed to concepts, and “possible worlds”. What I mean is those properties that characterize the subject, and which are not already implied by some other property. For example, being warm blooded is a property of man, but it is not an essential property of man, since man is a mammal (etc.), and “mammal” implies “warm-blooded”. An obvious essential property of man is having the faculty of reason, also having free will. I stop short of requiring that the identifications which constitute a theory have to be proven to the point of certainly; a fairly high standard of proof is necessary, to distinguish a theory from a hypothesis. And finally, an explanation about “subject”: this is basically shorthand for “the existents subsumed by a concept”.

Here are a couple of corollaries of this meta-theory. Because of the defining nature of “theory” – it is cognitive (it is created for a cognitive purpose) – theories inherit the economy requirements of concepts and their definitions. This derives various Occamite principles such as Aristotle’s “We may assume the superiority ceteris paribus of the demonstration which derives from fewer postulates or hypotheses”, and so on. “Grasping the nature of” an existent summarizes the Objectivist epistemology: it is a proper and objective relationship between a consciousness and reality. As a form of knowledge, there must be proper evidence for the claim, and a theory cannot be arbitrarily stipulated.

I would appreciate any criticism of this meta-theory directed at whether it does correctly describe what a theory is. It is irrelevant to me whether contemporary science teaching sees “theory” as a social construct. It is likewise irrelevant that most explanations of “theory” insist on adding stuff about repeated testing, standardized protocols or “testable”, since these are non-essential consequences of more basic concepts such as “knowledge”, “non-arbitrary”, or “probable” which the concept “theory” depends on. In other words, I’m trying to say what a theory is, and I am not trying to recapitulate what others have said about theories. I had hoped that How We Know would have a pre-packaged answer, but it does not seem to. Of course, alternative theories of theory important, since any claim has to be evaluated against reasonable alternatives.

Share this post


Link to post
Share on other sites

The first thing I thought of is how some cognitive development researchers, in terms of conceptual development, think of concepts as theories about the world. In order to have that theory of development, they need a theory about theories. Their aim isn't so much to say what the correct concepts are, but how children attain their conceptual understanding. The good researchers in this area deal with philosophy a lot, and sometimes it's hard to distinguish them from philosophers. What you write here is in general pretty consistent with that project. The set of highly probable propositions for a subject is the main idea for them I think. Of course, some of them are nativists, but I still think they're on the right track.

Your example of why warm-bloodedness as a nonessential property is a point I hadn't considered before. I always thought of nonessentials in terms of simply not explaining the most about the entity, just the opposite of how Rand thought of essentials. So when you say warm-bloodedbess this is basically essential to a supercategory (or subcategory) of the entity we want to understand, this sounds much easier to comprehend.

The only thing I'm unsure about is whether you actually distinguish theory from hypothesis. Would you explain that some more?

Share this post


Link to post
Share on other sites
12 hours ago, Eiuol said:

The only thing I'm unsure about is whether you actually distinguish theory from hypothesis. Would you explain that some more?

I don’t have an account of “hypothesis”, because the concept is outside my comfort zone. We mostly don’t deal in “hypotheses”, and my encounters with the concept come from allied areas of psychology and education. I am reasonably confident that a hypothesis is specifically about things that you do not know, and the standard of proof is a way to exclude hypotheses from being mistaken for theories, which was my intent. I think that a hypothesis is also specific to an individual, so it amounts to saying “this instance will have these properties”, on the grounds that “this instance” is a concrete example of a theoretical concept. If that is correct, then a hypothesis is just a prediction applied to an instance. Alternatively, a hypothesis may be even less grounded in reason and could just be a conjecture. But, as I say, my profession generally doesn’t talk that way, so I’m not comfortable with the term.

Share this post


Link to post
Share on other sites

Maybe this makes sense to you (from Mario Bunge, Philosophical Dictionary, Enlarged Edition, 2003): 

THEORY Hypothetico-deductive system: i.e., a system composed of a set of assumptions and their logical consequences. In other words, every formula of a theory is either an assumption or a valid consequence (theorem) of one or more assumptions of it […]. Again: a theory is a set of propositions closed under deduction (i.e., including all the logical consequences of the axioms).

Most people, even some philosophers, confuse theory with ↑hypothesis. This is a mistake, because a theory is not a single proposition but an infinite set of propositions. Therefore it is far more difficult to confirm or to falsify than a single hypothesis. (Analogy: a net is stronger than either of its component threads, hence harder to make and to rip.) Another serious confusion is that between theories and &languages. This is a mistake because theories make assertions, whereas languages are neutral. The mistake is part of formalism, the mathematical component of nominalism.

A theory may refer to objects of any kind, nondescript or well-defined, conceptual or concrete, and its assumptions may be true, partially true, false, or neither. The condition of logical deducibility from the initial assumptions confers formal (syntactical) unity upon the theory. This allows one to treat theories as (complex) individuals. These individuals possess emergent properties that none of their components (propositions) possess, such as consistency (noncontradiction).

Example 1: Set theory, graph theory, and Boolean algebra are abstract (uninterpreted) theories. Example 2: Number theory, Euclidean geometry, and the infinitesimal calculus are interpreted mathematical theories. Example 3: Classical mechanics, the theory of natural selection, and neoclassical microeconomics are factual theories. Nonexample: The assumptions “All As are Bs,” and “All Cs are Ds,” where the predicates A, B, C, and D are not inter-definable, do not constitute a system, hence do not generate a hypothetico-deductive system. Indeed, no consequences derive jointly from them.

HYPOTHESIS Educated guess. A statement that embraces more than the data that suggest or confirm it. All the empirical generalizations and law-statements, even the well-confirmed ones, are hypotheses. Thus, human knowledge is largely hypothetical. However, not all hypotheses are equally plausible: whereas some are proffered as tentative, others are regarded as very close to total truth, and still others as final: plausibility. Examples of definitive truths that started out as tentative hypotheses: “The universe has evolved,” “There are force fields,” “RNA takes part in protein synthesis,” and “Individual decision making is located in the prefrontal cortex.”

Edited by AlexL

Share this post


Link to post
Share on other sites

I make a distinction between "theory" and "model" when I think it is necessary.  A theory explains certain facts in terms of others.  A model predicts certain facts from others.  So, for example, chemistry is a theory and a model, because it contains explanatory relationships as well as predictive methods.  Quantum mechanics is a pure model, devoid of (generally accepted) metaphysics.  (But QM plus the "many worlds" interpretation would be a theory and a model.)  Our understanding of much psychology is theory, rather than model, explanatory without much predictiveness.

Re: Hypothesis and theory.  I tend to think of them as points on a continuum.  A hypothesis is merely a set of explanatory relationships suggested by the facts.  A hypothesis becomes a theory when its consequences have been explored without finding contradictions.  Go further and the theory becomes a certainty, a fact in itself.

Share this post


Link to post
Share on other sites

I do understand the words: Bunge's position this is more or less the approach to theory which I am arguing against. For example, a theory is not a deductive system, it is an inductive system. Deduction may play some role in validating a theory: a theory does not contain logical consequences, it implies them. A theory does not contain assumption (conjectures), it contains truths which are the generalized product of research. The idea that “theory is assumptions” is the credo of that approach to theory which allows arbitrary statements to be put out there, and we don’t do that. A theory is not an infinite set of propositions (if it were, nobody could know a theory, probably not a distressing conclusion to folks of that ilk). An infinite set of highly specific propositions could follow from the few propositions that comprise a theory (the problem is that there is not an infinite set of existents, so infinitely many of those “propositions” would be based on an invalid premise “If only it were true that X152,349 existed”).

I don’t accept his claim that a theory is harder to confirm or falsify than a hypothesis, largely because I don’t know what it would even mean to be “hard” to confirm or falsify, and making bad analogies to nets does not clarify the logic of the statement. I would counter that it is vastly easier to refute a theory than to refute a hypothesis (if we take “hypothesis” to be about a very specific instance, within the domain covered by a theory). A theory entails very many specific hypotheses and if any of them are false, the theory is false (100,000 true values is highly unlikely). A hypothesis entails basically one of two logically possible outcome, where falsification is comparatively very unlikely –50-50). The characterization of hypothesis as “A statement that embraces more than the data that suggest or confirm it” fails to distinguish theory from hypothesis (this definition is also true of “theory”). Again,  “All the empirical generalizations and law-statements, even the well-confirmed ones, are hypotheses” is also true of theory. The guy seems to be confused about standard of proof vs. cognitive function. 

This is a useful source, where I can show that I'm not arguing against a straw man.

Share this post


Link to post
Share on other sites
On 4/7/2018 at 2:43 AM, DavidOdden said:

a theory is not a deductive system [...]

The moment a body of specialized knowledge reaches the status of being a theory , the inductive phase is already far behind. Indeed, the observation of facts, the elaboration and integration into concepts, the identification of causal connection and (at least partial) attempts to quantitatively describe the facts precedes the theory stage. A fully elaborated theory is a body of knowledge which is already hierarchically structured, with clearly identified presuppositions which logically and mathematically implied consequences.

 Example: electromagnetism. The body of knowledge starts with simple observed facts, like entities which, under certain circumstances, attract or repel each other. Then it is hypothesized that the entities are electrically charged, and this causes the observed effects. Then comes Coulomb (and Cavendish, ca. 1780) and quantitatively describes the phenomenon. On the other hand, there are magnets which are observed to interact, and the attempt to describe the magnetic interaction similarly to the electric ones fails completely, which indicates they differ radically in their nature. Then it is observed that the wires with moving electric charges (currents) exhibit magnetic properties (Ampère, ca.  1820). The existence of electric and magnetic fields is hypothesized. Then comes the realization that they are not independent. Finally comes Maxwell and unifies both (ca. 1860) with his equations. With this, the theory of electricity and magnetism, “electrodynamics”, is, practically, fully elaborated. Logical and mathematical consequences – deductions - are then studied, tentatively applied to old and new phenomena, and thus solidly confirmed.

 The most spectacular mathematical consequence was the prediction – deduction - of self-sustaining, propagating electromagnetic waves, with a speed closed to the known speed of light. Maxwell hypothesized, then Hertz confirmed (1887) that the light is an electromagnetic wave.

 Thus, classical electrodynamics is a theory in Bunge’s sense, that is a body of knowledge presently organized as a hypothetico-deductive  system: the hypothesis are the inductively reached assumption of the existence of electric charges, of electric and magnetic fields, and of the Maxwell equations to unify the whole ménagerie :-). Next, all the observed phenomena of the corresponding nature were explained – deduced - using the assumptions (Coulomb, Ampère, and other empirical laws), as well as new ones discovered (e.m. waves, for eyample).

The Maxwell’s theory of electromagnetism can be applied to an unlimited number of particular circumstances, differing by the relative arrangements of charges and currents, with different environments. IOW, the theory of classical electrodynamics has an potentially unlimited number of consequences.

Though your (first?) impression of Bunge was negative, I still think you will find him interesting at least in part: he is a very good and profound physicist, and, as a professional philosopher, completely different from what one usually sees nowadays.

Therefore, I recommend you warmly:  Mario Bunge, Philosophy of Physics, 1973 (found in any scientific library, I hope). It is mainly for physicists, but you might be interested in his approach which he explains in detail.

Edited by AlexL

Share this post


Link to post
Share on other sites

I think it may well be that "theory" in physics is different from "theory" in other sciences, owing to the fact that physics is significantly advanced. Not being a physicist, I don't have a specific exemplar of "theory" to contemplate. A lot of what I vaguely know about modern physical theories is not about anything empirical, it's about alternative mathematical descriptions. I suppose I need to find a geologist or virologist, to see what examples of "theory" are in those disciplines.

Share this post


Link to post
Share on other sites
22 hours ago, DavidOdden said:

I think it may well be that "theory" in physics is different from "theory" in other sciences, owing to the fact that physics is significantly advanced. Not being a physicist, I don't have a specific exemplar of "theory" to contemplate.

I would rather say that the meaning of “theory” is the same, but, as I argued previously, not any chunk of knowledge, even obtained by the scientific method, deserves to be called “theory”. As you noted, it has to be advanced enough to be a theory. But among natural sciences, physics (and other exact sciences) are not the only ones to be or to contain theories.

22 hours ago, DavidOdden said:

I suppose I need to find a geologist or virologist, to see what examples of "theory" are in those disciplines.

It so happens :-) that for geology the same M. Bunge makes, in a different work, some observations which might interest you, for example his assertion that the tectonic plates explanation of “the present configuration of mountain ranges and ocean bases, as well as many earthquakes” is a theory. The work is his multi-volume “Treatise of Basic Philosophy” (vol. 7, part I, p. 231), available in bigger scientific libraries. Or just PM me for the relevant pages in electronic format.

Share this post


Link to post
Share on other sites

A scientific theory is a collection of assumptions about some field, which, given a model of an experiment in that field, produces a prediction about the outcome of that experiment.

Share this post


Link to post
Share on other sites
1 hour ago, SpookyKitty said:

A scientific theory is [...]

A theory is not about experiments and their outcomes, but about causal connections.

Edited by AlexL
Grammar

Share this post


Link to post
Share on other sites
2 hours ago, AlexL said:

A theory is not about experiments and their outcomes, but about causal connections.

If this were true, then Newtonian Physics is not a scientific theory. Newtonian dynamics is governed by the second law, which has nothing at all to do with causes and effects.

Share this post


Link to post
Share on other sites
3 hours ago, SpookyKitty said:

If this were true, then Newtonian Physics is not a scientific theory. Newtonian dynamics is governed by the second law, which has nothing at all to do with causes and effects.

Isn't (net) force the cause and acceleration the effect?

Share this post


Link to post
Share on other sites
28 minutes ago, human_murda said:

Isn't (net) force the cause and acceleration the effect?

No. A cause must precede its effect. The acceleration and force are always simultaneous.

Share this post


Link to post
Share on other sites
36 minutes ago, SpookyKitty said:

No. A cause must precede its effect. The acceleration and force are always simultaneous.

Force produces a change in the state of motion of the object (the change in the state of motion occurs after the force is applied). Acceleration is a measure of that change across time. Just because the numerical values of acceleration and force are simultaneously defined (and can be mathematically obtained from each other) doesn't mean one doesn't cause the other.

Share this post


Link to post
Share on other sites
1 minute ago, human_murda said:

Force produces a change in the state of motion of the object (the change in the state of motion occurs after the force is applied).

No. The acceleration is produced at the exact moment that the force is applied. Not a finite amount of time afterwards.

Quote

Acceleration is a measure of that change across time.

Acceleration is a measure of the instantaneous rate of change in velocity over time.

Quote

Just because the numerical values of acceleration and force are simultaneously defined (and can be mathematically obtained from each other) doesn't mean one doesn't cause the other.

This is a misrepresentation of my argument. A cause must precede its effect chronologically. Because forces produce accelerations instantaneously, one cannot cause the other.

Share this post


Link to post
Share on other sites
24 minutes ago, SpookyKitty said:

Acceleration is a measure of the instantaneous rate of change in velocity over time.

All derivatives (including acceleration) are defined around an infinitesimally small interval around a particular value in the function domain. Not only that, you need a concept of ordering. In the case of time derivatives, you need time ordering (before and after). Without it, you'll get the wrong sign of acceleration. For example, without the correct time ordering (which keeps track of the changes brought about by forces), you'll get an acceleration vector that is in the opposite direction of force (which is physically untrue). You definitely need a concept of before and after in the case of a derivative. Otherwise, the law wouldn't work. These things are implicit in the workings of calculus.

(To say that acceleration is instantaneous is true. To say that it involves no concept of before and after is context dropping. Such concepts are implicit in the meaning of instantaneous).

Edited by human_murda

Share this post


Link to post
Share on other sites
18 minutes ago, human_murda said:

All derivatives (including acceleration) are defined around an infinitesimally small interval around a particular value in the function domain. Not only that, you need a concept of ordering. In the case of time derivatives, you need time ordering (before and after). Without it, you'll get the wrong sign of acceleration. For example, without the correct time ordering (which keeps track of the changes brought about by forces), you'll get an acceleration vector that is in the opposite direction of force (which is physically untrue). You definitely need a concept of before and after in the case of a derivative. Otherwise, the law wouldn't work. These things are implicit in the workings of calculus.

(To say that acceleration is instantaneous is true. To say that it involves no concept of before and after is context dropping. Such concepts are implicit in the meaning of instantaneous).

I never said that.

Share this post


Link to post
Share on other sites
6 hours ago, SpookyKitty said:

Newtonian dynamics is governed by the second law, which has nothing at all to do with causes and effects […]

A cause must precede its effect. The acceleration and force are always simultaneous […]

Because forces produce accelerations instantaneously, one cannot cause the other […]

The acceleration is produced at the exact moment that the force is applied. Not a finite amount of time afterwards […]

What exactly troubles you?

1. Why should there be a finite amount of time between the moment the force is applied to a mass and the moment its velocity starts to change? How big do you expect this delay to be and what it is needed for, physically?

2. You say the Newtonian mechanics is not causal. Is the special relativity (which is capable of accounting for finite speed of propagation of interactions) better in this respect?

Note: I define cause by the fact that if one suppresses it, the effect disappears – if you suppress force, there is no acceleration any more. I do not define causality by temporal antecedence. How do you define cause?

Edited by AlexL

Share this post


Link to post
Share on other sites

An entity causes its actions; which actions depend on the nature of the entity and the circumstances surrounding it.  Those circumstances may change, in which case the entity may act differently than it did before the change. When that happens, one can say that the change is a "cause" and the difference in action an "effect" and thereby speak of cause and effect.  This way of speaking must be used carefully, because it leads to the, among other things, the confusion displayed above.  However, using that language, it should be clear that there is no requirement that an effect occur at the same time as its cause or at some later time; when an effect occurs will depend on the nature of the entity acting.

Share this post


Link to post
Share on other sites
16 hours ago, SpookyKitty said:

A scientific theory is a collection of assumptions about some field, which, given a model of an experiment in that field, produces a prediction about the outcome of that experiment.

When you say “collection of assumptions”, is this intended to refer to refer to something different from my “system of identifications”? Typically people see “assumptions” as being closer to the “possible” end of the evidentiary scale, so is that what you have in mind. Your theory of “theory” relies on the concept “model”, and I do not know what a model is for you. I assume that in using “experiment”, you’re referring to any kind of observation.

To be concrete, can you illustrate your characterization of theory by presenting a simple theory of “atomic nucleus”? I have not thought about this, myself, and I’m not a physicist so it is certainly a very amateur theory, but I would say that a nucleus is the dense center of an atom, composed of protons and neutrons which are held together by the nuclear force. I would also say something about the relationship between the number of protons in a nucleus and what an atom “does”, but that’s above my pay grade. The two things I’m most interested in are (a) whether I’ve left out important assumptions that your theory would include (I'm addressing the question of essentiality here), and (b) what are one or two predictions of a/your theory of the nucleus? The point is that I’d like to understand how you make “prediction” part of what it means to be a theory. In my account, predictions derive from something different (knowledge of a thing’s nature means knowing what it does, so this is separate from the concept “theory”).

Share this post


Link to post
Share on other sites
On 5/1/2018 at 6:05 AM, human_murda said:

Anyway, the original statement of Newton's second law is:

4ae898a014e2512dcff748dbe6225543f86e3d84

This describes the effect after the cause (impulse).

No. It simply doesn't.

On 5/1/2018 at 9:15 AM, AlexL said:

What exactly troubles you?

1. Why should there be a finite amount of time between the moment the force is applied to a mass and the moment its velocity starts to change? How big do you expect this delay to be and what it is needed for, physically?

There shouldn't be one.

Quote

2. You say the Newtonian mechanics is not causal. Is the special relativity (which is capable of accounting for finite speed of propagation of interactions) better in this respect?

Yes. Definitely. In SR, forces are entirely local, and so there is no instantaneous action at a distance.

Quote

Note: I define cause by the fact that if one suppresses it, the effect disappears – if you suppress force, there is no acceleration any more. I do not define causality by temporal antecedence. How do you define cause?

So if I close my eyes, then the sun ceases to shine? Your definition of causality is absurd.

 

On 5/1/2018 at 10:35 AM, DavidOdden said:

When you say “collection of assumptions”, is this intended to refer to refer to something different from my “system of identifications”? Typically people see “assumptions” as being closer to the “possible” end of the evidentiary scale, so is that what you have in mind. Your theory of “theory” relies on the concept “model”, and I do not know what a model is for you. I assume that in using “experiment”, you’re referring to any kind of observation. 

By "assumptions" I mean statements which help simplify the problems in the field. No feasible scientific theory can take into account literally everything which could affect the outcome of the experiment. Furthermore, the role of these assumptions (fundamentally) is to restrict the range of possible models, so as to facilitate computation of outcomes of experiments or to make it is easier to search the space of all models quickly and/or efficiently and also to evaluate them.

Quote

To be concrete, can you illustrate your characterization of theory by presenting a simple theory of “atomic nucleus”? I have not thought about this, myself, and I’m not a physicist so it is certainly a very amateur theory, but I would say that a nucleus is the dense center of an atom, composed of protons and neutrons which are held together by the nuclear force. I would also say something about the relationship between the number of protons in a nucleus and what an atom “does”, but that’s above my pay grade. The two things I’m most interested in are (a) whether I’ve left out important assumptions that your theory would include (I'm addressing the question of essentiality here), and (b) what are one or two predictions of a/your theory of the nucleus? The point is that I’d like to understand how you make “prediction” part of what it means to be a theory. In my account, predictions derive from something different (knowledge of a thing’s nature means knowing what it does, so this is separate from the concept “theory”).

I think I can come up with an even simpler domain over which we can explore the concept of scientific theory. Let's say that you're investigating an unknown (possibly alien) language. You have some documents written in that language and one of them says the following:

"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum."

And your job is to figure out what are the possible words of this language. (Actually, I'm currently working on an AI which solves this problem). One possible theory is that the words in a language consists of a small set of syllables which can be combined to form the words in that language. (This is our first assumption). So a model that this theory produces, for example, could say that some of the syllables of this language are:

lo, rem, ip, sum, do, sit, a, met,....

And so, it would allow us to predict that we can have words like "lorem" and "ipsum" which are actually in the document but also we could have words like "sumdosit" and "aremdo" which aren't.

Another possible model is that the only syllables are:

a,b,c,d,e,f,g,h,i,k,l,m,n,o,p,q,r,s,t,u,v,x,y,z.

So this model also predicts words like "lorem" and "ipsum", but it also predicts words that the first model doesn't, for example, "tstrqpomlk" and "defghrp".

Which model do you think is a better explanation of the data? An essential component of a theory is a method for evaluating its models. So for our theory we could require that good models first of all assign very high probability to the data that we actually observe. This is so that we can make very precise and correct predictions about the data. By this criterion, the first model is clearly superior to the second.

On the other hand, suppose that we come up with the following third model:

"lorem, ipsum, dolor, sit, amet,..."

where the syllables of the language are precisely the words we observe in the document. This model assigns extremely high probability to the data (In fact, you can't possibly do better). But it seems to sort of cheat because it is overfitting the data. So we could introduce a third assumption about what makes a good model, that is, that a good model minimizes the ratio (#of characters needed to specify the syllables in the model)/(#of characters in the data). This would prevent overfitting.

The algorithm that I designed uses these assumptions to produce models of arbitrary languages which are more like the highly precise and very simple first model and less like the imprecise but simple second model and also less like the extremely precise but absurdly complicated third model.

Share this post


Link to post
Share on other sites
12 hours ago, SpookyKitty said:

There shouldn't be one [that is - a finite amount of time between the moment the force is applied to a mass and the moment its velocity starts to change]

OK, but then, if in the Newton‘s Second law a = F/m , F is the force at exactly the point (x,y,z) where the point mass m is located and at exactly the moment t, then the acceleration at that moment is indeed given by the above formula: a(t) = F(x,y,z; t ) / m, with no delay. Do you agree?

With this - local - value of F, is the Newton Law better (at least for moderate velocities) and the Newton’s theory – really a true theory (in your sense) and an excellent approximation of reality?

Can you now view a(t) as the effect and F(x,y,z; t ) as the cause? If not, why not?

Edited by AlexL

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

×