What is science? Its method? Its limits?

[Hegemonicretribution]

Just as the title suggests. What do you consider science to be? How is it distinct from all non-scientific approaches.

How do scientists function? What are experiments? Do they proove/disproove or simply lend support/ confirmation to a theory? To what extent would confirmation be given?

Is explanation, prediction, or both the aim of science? Which is more important?

What is the most we can hope to achieve with science? Does induction cause us a problem, or theory selection? How do we choose between theories? Are scientific descriptions true pictures of reality, or simply instrumentally useful?

With current funding practices how much belief can we place in scientific findings? Given a collectivist approach to science would these questions change? What problems would be solved/generated.

.................................................. ...................

I know a lot of people point to science as a higher form of knowledge. Terms such as 'proof' or 'refutation' are thrown around here a lot. Also philosophy is often hated on. What I want to get at is what exactly is this science upon which we all depend? I would like to tackle some of the big questions here and see if we cannot achieve a more reasonable understanding of what science is/does/can do/should do.

[mikelepore]

My definition is: Science is recognition that there is some regularity in the universe, and the efforts to determine under what conditions the common patterns reappear. If it's possible to experiment with something, you do. If you can't do experiments, at least you try to name the classes of things and list their properties.

[Hegemonicretribution]

What form does this regularity take? Is it independent from theoretical observation?

I see that you did not specify a uniformity of nature, which is fortunate, but I do not know exactly what is being proposed if not this. How do we determine what conditions are the same as those which previously expressed some form of regularity? What if there is a break from this regularity, do we consider the regularity to be imaginery, or do we consider the conditions to be somehow different?

As for clasifying things which we cannot experiment upon, why? Why list their properties? If you can not even potentially verify this, then what is the point? If you can be aware of their properties surely they can be experimented upon?

I know that I am being a little contrary here, but philosophy of science is largely where my passion lies, and I feel that it is perhaps the only live branch of philosophy at present. I am trying to establish, or at least gauge roughly how most people conceive of science, and why they place a certain degree of faith in it in spite of the problems that I raise here, or a more general ignorance as to what it is. Much appreciated, and I look foward to further responses.

[black magick hustla]

This question is as nonsensical as asking oneself what is beauty.

Such words have no meaning outside their respective language games. Do you want the anthropological take on it? I.e. how science is perceived by different communities? The realist interpreteation of science?

I think most people view science as the practice of the scientific method - which implies using empirical data to extrapolate some sort of paradigm that could yield reasonable predictions. Anything beyond that is platonism, I think.

[black magick hustla]

I like philosophy of science. I am a physics major and is always something interesting to discuss. A lot of my classmates and coworkers see themselves as some sort of ivory tower of truth. What I like about science is that it deals with truth-functions i.e. statements that can either be true or false. Is the cat on the mat? Is the orbit of the earth eliptical? However the scientific community, as marx said, have "traditions weighting like a nightmare on their brains" like any other community and they can be subject to many superstitions too - like too much faith on mathematical models, for example.

[Iowa656]

Science the way to model the universe based on evidence.

I make the statement A (This could be any claim, "truthful" or not, ie the world is flat, atoms are made of hadrons and elections etc)

I find evidence to prove or disprove my statement. (Any form of data is collected, experiments, real life observations, ie I see that the world has no "end", thus it cannot be flat).

I modify my claim and improve it based on the evidence produced. (The model presented is improved or modified to align with the real life data produced. The earth is not flat, thus it must be a different shape, namely spherical.)

I use this to predict the outcome of other events in the universe (ie, other planets are round also).

Prediction is tested based on new evidence. And repeat.

[mikelepore]

Science discovers the kind of regularities where a brief statement encompasses many cases. When you say the little phrase "conservation of momentum" you just explained the collision of billiard balls, human motions in dancing and ice skating, the kickback of a gun, the orbits of the planets, and more. (That's an example where experimentation made the generality known.)

Example of knowledge that can be gathered by observation and classification only, even if the means of experimentation to test hypotheses were not yet discovered: While all animals with four chambered hearts are alike in that all of them are warm blooded, they are not all alike in that those with hair produce milk, while those with feathers do not. Animals with three or two chambered hearts are all alike in that they are all cold blooded and never have feathers, but they are not all alike in that, while some animals with three chambered hearts have scales, and others have skin, all animals with two chambered hearts have scales.

Such facts are regular patterns in the world. This regularity is the subject of science. Science seeks to find out about it.

[mikelepore]

Quoted from the essay "On Nature" by John Stuart Mill (1874)

"As the nature of any given thing is the aggregate of its powers and properties, so Nature in the abstract is the aggregate of the powers and properties of all things. Nature means the sum of all phenomena, together with the causes which produce them; including not only all that happens, but all that is capable of happening; the unused capabilities of causes being as much a part of the idea of Nature as those which take effect. Since all phenomena which have been sufficiently examined are found to take place with regularity, each having certain fixed conditions, positive and negative, on the occurrence of which it invariably happens, mankind have been able to ascertain, either by direct observation or by reasoning processes grounded on it, the conditions of the occurrence of many phenomena; and the progress of science mainly consists in ascertaining those conditions. When discovered they can be expressed in general propositions, which are called laws of the particular phenomenon, and also, more generally, Laws of Nature."

http://www.marxists.org/reference/archive/mill-john-stuart/1874/nature.htm

http://www.lancs.ac.uk/users/philosophy/texts/mill_on.htm

[mikelepore]

Copied from "Cosmos" by Carl Sagan, Chapter 3:

"If we lived on a planet where nothing ever changed, there would be little to do. There would be nothing to figure out. There would be no impetus for science. And if we lived in an unpredictable world, where things changed in random or very complex ways, we would not be able to figure things out. Again, there would be no such thing as science. But we live in an in-between universe, where things change, but according to patterns, rules, or, as we call them, laws of nature. If I throw a stick up in the air, it always falls down. If the sun sets in the west, it always rises again the next morning in the east. And so it becomes possible to figure things out. We can do science, and with it we can improve our lives."

[Hegemonicretribution]

Marmot, whilst I find what you have said is interesting, I am concerned with the boundaries of science and its application. As such, treating all of science as a language game does not answer these questions. I am not too interested in what x group of people think about science and how they use the words particular to them....I am trying to establish logical limits on a scientific method, distinguish this method as it is from how it is conceived of, and illustrate problems. I guess I am talking about science as a 'philosopher of science' as would be evidenced by the questions asked. Consequently I am not asking about any particular view of science, I am aware of many of these, what I am looking for is a good desription of what science actually is. That is not according to a realist, or instrumentalist, or hypothetico-deductivist etc...I know what these people say on the matter. I am more concerned with a reasonable and accessible account which could underpin the higher degree of 'belief' we place in whatever it is that we call science.

Iowa656:
I find evidence to prove or disprove my statement. (Any form of data is collected, experiments, real life observations, ie I see that the world has no "end", thus it cannot be flat).
How does evidence proove a statement? I can see how it may well disproove it, but proof does not appear possible here.

If each new piece of evidence offers proof as a matter of degrees, then to what extent does a piece of evidence x proove a hypothesis y?

You go on to say that you modify your hypothesis; does piece of evidence x equally add support to hypothesis z as it did to the original hypothesis y?

Evidence 'produced' also sounds problematic? I take it that this evidence is produced in accordance with a hypothesis. Given the theoretical nature of the observations made in such a case, would you not require another example? The flat earth hypothesis relied only upon generally accepted hypotheses about geometrical shapes, and our ability to observe the world. If you are producing evidence in accordance with a hypothesis then you are relying upon hypotheses contingent upon the general acceptance of your claims. In other words in undermining your theory you also undermine your evidence. I am not saying this is what you necessarily intended, or even that it is fatl, but I am highlighting a problem.

miklepore


(That's an example where experimentation made the generality known.) This is true so long as a counter example does not arise. If it does then the generality was not known, it was assumed to be one way on the back of a certain ammount of evidence, but the over all account would be lacking. Perhaps not problem in your example, but it is generally for scientific method. Without assuming a uniformity in nature you cannot expand to the general without finding something that holds in common between everthing you wish to construe as part of 'the general'.



Example of knowledge that can be gathered by observation and classification only, even if the means of experimentation to test hypotheses were not yet discovered

The hypothesis there is that patterns resulting from past observations will hold in future observations. The means to test this hypothesis is simply observing new types of animals; ones which have not previously been observed. The more animals that are in accordance with the hypothesis, the better supported it is. Any animals which break with the hypothesis cause a serious problem for the overall theory.
I do not see what the hypothesis here actually is, if not the clasification and observation itself.....if it is the sum of observations, then the observations provide a means to test the hypothesis. If you want to suggest that the hypothesis is something more than this clasification, then I challenge you to assert what it is. Essentiallym, in offering a clasification you imply a particular hypothesis which meets with these clasifications. If by hypothesis you mean 'why it is that such divisions exist' then I see where you are coming from, but again the hypothesis involved in clasifying observations still implies the continuing regularity of observations as necessary for acceptance of the overall theory.


Such facts are regular patterns in the world. This regularity is the subject of science. Science seeks to find out about it.

I accept that there is some degree of regularity, and that science involves itself with explaining/dicovering this regularity, but this is not the whole picture. It also seeks to explain irregularity. This has not been accounted for. Neither has the manner in which science 'seeks to find out about' regularity. It is this with which I am more concerned. 'Social sciences' try to find out about, predict and explain certain regularities whether perceived or real, so why is this different from scientific method?

More generally, is the illustration, explanation or prediction of said regularity more important? I assume they are all inter-connected?

Perhaps most importantly, given that I may not of been clear about what you previously said (but I am very clear on this); are the patterns percieved in reality objective features of the world, or theoretically dependent explanations pf perceived regularity?

As for the JS MIll quotation, I wil not dismiss them simply as liberal, or something similar because they do have some good ideas. However, and I hope you will not feel that I am avoiding the issue, Mill was writing before philosophy of science or the modern scientific method really took off. As such it would be unfair to attack them for failing to account for later criticisms to which they never got to reply.

The passage by Sagan I will not take issue with either, as I do not deny something approaching regularity in the world, or the possibility of science. What I am concerned about is what science is, not whether or not it is possible. Although you have not said much to which I would object, you do leave me asking a number of questions, and it is these questions which I find more interesting because it is about this point that everything becomes somewhat murkier. Again thank you all for responding.

[mikelepore]

Perhaps most importantly, given that I may not of been clear about what you previously said (but I am very clear on this); are the patterns percieved in reality objective features of the world, or theoretically dependent explanations pf perceived regularity?

In science, The things about which we think and communicate are objective features of the world. But every symbolic notation to permit thinking about them and communicating about them depends on the culture and individual idiosyncracy.

[mikelepore]

More generally, is the illustration, explanation or prediction of said regularity more important? I assume they are all inter-connected?

"More important" isn't a factual matter. It's a description of someone's state of mind. Perhaps an architect only cares about whether a building will stand up. There's nothing wrong with that either.

[mikelepore]

The hypothesis there is that patterns resulting from past observations will hold in future observations. The means to test this hypothesis is simply observing new types of animals; ones which have not previously been observed. The more animals that are in accordance with the hypothesis, the better supported it is. Any animals which break with the hypothesis cause a serious problem for the overall theory.
I do not see what the hypothesis here actually is, if not the clasification and observation itself.....if it is the sum of observations, then the observations provide a means to test the hypothesis. If you want to suggest that the hypothesis is something more than this clasification, then I challenge you to assert what it is. Essentiallym, in offering a clasification you imply a particular hypothesis which meets with these clasifications. If by hypothesis you mean 'why it is that such divisions exist' then I see where you are coming from, but again the hypothesis involved in clasifying observations still implies the continuing regularity of observations as necessary for acceptance of the overall theory.

My point is that doesn't have to be any hypothesis at all for a task to be a useful part of the process of science. Perhaps you have heard about Ernest Rutherford's remark, "All science is either physics or stamp collecting." On some days you just sort through a heap of raw data and try to put things into envelopes that are labelled in a way that will help other people come back to them in the future. On other days, you work with hypotheses and models and principles. Both are necessary parts of science. There are a couple people at NASA whose whole job is to write lines of software that will make the Mars rovers go around a rock and not bump into it. They're doing science too. The understanding, which will eventually consist of hypotheses that people will look back on, will emerge from the whole project. But a lot of the daily work is, as Rutherford said, like stamp collecting.


I accept that there is some degree of regularity, and that science involves itself with explaining/dicovering this regularity, but this is not the whole picture. It also seeks to explain irregularity. This has not been accounted for.

Any irregularity becomes just another regularity as soon as you determine its extent. If the measurement values jump around erratically, how much deviation is there? Did tweaking any parameters cause the mean or standard deviation to move? Now you have another regularity to report. But since I started out mentioning the taxonomy of animals, I'll stick with that. Irregularity appears in the form of an exception to an earlier block on the chart. First you may say that fish have bones. Then you correct the statement for a subclass of cartilaginous fish that don't have bones, which probably becomes the favorite specialization of some other researchers. We're not losing the pattern recognition that we have for understanding nature, but just filling out the picture.



Neither has the manner in which science 'seeks to find out about' regularity. It is this with which I am more concerned.

That's a procedural issue, which you can read about in an encyclopedia or something.


'Social sciences' try to find out about, predict and explain certain regularities whether perceived or real, so why is this different from scientific method?

Social sciences follow the scientific method to the extent that the tools are available. We wouldn't allow sociologists to keep human societies locked in glass containers for them to conduct controlled experiments on, so that tool isn't available. Identifying correlations is less revealing, but at least its a tool that is easily available.

[mikelepore]

How does evidence proove a statement? I can see how it may well disproove it, but proof does not appear possible here.

Now that I've got Rutherford on my mind ... it is said that his gold foil experiment proved something very clear. He caused a beam of alpha particles, which have a positive charge, to pass through a thin foil. The vast majority of the particles went straight through, but a small percentage were deflected by large angles, some even bouncing back. His conclusion had to be that atoms are not filled with stuff throughout their volume, as the old "plum pudding" model of the atom had supposed, but, rather, the atom is mostly empty space, with most of its mass contained within a small nucleus. Furthermore, it had to be that it's the positive charge in the atom that resides in that nucleus, and not the negative charge, to account for the repulsive deflections. By the peculiarity of grammar, we may phrase it to say that he disproved one idea, or we may phrase it to say that he proved a different idea.

[Iowa656]

How does evidence proove a statement? I can see how it may well disproove it, but proof does not appear possible here

That is indeed correct, it's a lot easier to disprove a statement that it is to prove one. Whatever my statement is, one piece of evidence can disprove it, but an "infinite" amount of evidence is needed for definite proof.

It depends on how you are defining proof. From a mathematical view you must prove a statement true for all value in order for it to be true, any value for which it is not true means the statement is false.

For example I make the statement that the square of (x+1) is equal to the square of x plus two x plus one. This can, by expanding the term (x+1) squared, be proven true for all x. So we accept this as a fact. In contrast I then make the statement x squared minus one equals zero and I demonstrate how if x=1 the statement hold true, I might naively I suggest the statement is true for all x. Which evidently it is not.




If each new piece of evidence offers proof as a matter of degrees, then to what extent does a piece of evidence x proove a hypothesis y?

That depends on how your mind works.

Mathematically a statement must be 100% true for it to be accepted.

In many other conditions however a less conclusive proof can be accepted.

For example, say a specimen of DNA was found on a murder weapon. Then it was checked with a sample of my DNA. A scientist might then conclude that there is a 99.99% chance it is my DNA, thus present the case that I committed the murder. If you are a judge is 99.99% chance sufficient to convict me? If it is then one in 10,000 cases are statistically likely to be a false conviction.



You go on to say that you modify your hypothesis; does piece of evidence x equally add support to hypothesis z as it did to the original hypothesis y?

Hypothesis z is a extension of hypothesis y, usually including hypothesis y in it.

For example, my first hypothesis, hypothesis y, is that all matter is made out of atoms. I conduct many experiments and they present the fact this is true. I can mix chemicals and produce different elements. I can break down matter into atoms. I then accept the hypothesis. Next, and in fact hundreds of years latter, hypothesis z is that atoms are made of hadrons and electrons. I do experiments in a particle accelerators, using radiation or using charged particles that demonstrate this to be true. I then accept this hypothesis z. Does this mean that hypothesis y is false? Not really; it is true, yet is insufficient. Matter is made of atoms and atoms are made of hadrons and electrons. This is pretty much how science works, apart from occasionally new evidence shows the first hypothesis to be completely false; for example, hypothesis 1, earth is stationary everything else moves, evidence moving stars, sun and planets in sky, hypothesis 2, earth in fact moves round sun, evidence, false predictions in orbits, evidence of moons orbiting other planets, etc.



Evidence 'produced' also sounds problematic? I take it that this evidence is produced in accordance with a hypothesis.

Evidence isn't really produced, more accurately would be a description that is it observed. Example, I drop a ball, I observe it, it falls to the ground, I repeat, that is my experiment. I have produced data; all objects fall to earth.



Given the theoretical nature of the observations made in such a case, would you not require another example?

Oh indeed, no one would claim X does this, so everything does this.



The flat earth hypothesis relied only upon generally accepted hypotheses about geometrical shapes, and our ability to observe the world. If you are producing evidence in accordance with a hypothesis then you are relying upon hypotheses contingent upon the general acceptance of your claims. In other words in undermining your theory you also undermine your evidence. I am not saying this is what you necessarily intended, or even that it is fatl, but I am highlighting a problem.

Usually in science it is the other way round. The evidence is this, so the theory is built upon it. In the flat earth example, the evidence was that earth seemed flat to our perceptions and that ships rarely returned form long journeys, thus the theory was that the earth was flat. New evidence became apparent and flaws were found in this first theory, and such it was improved.

If my evidence if damming enough to show the claim is true for all such situations as you can possibly demonstrate, but a chance of being false if repeated for "infinite" trials then can the hypothesis be accepted as true? How much repetition is required? For example, again, I claim that objects will fall to the earth, is it possible to test this with each and every object? Of course not, however evidence suggests that it is true. If suddenly an object was mysteriously found that has mass and did not obey the laws of gravity then all that we know would be questioned. One counter example is enough to debunk the whole theory.

Have you heard of Fermat's last theorem? I recommend you read a bit about it. It's an incredible mathematical theorem that was presumed true for 200 years. Only then was it proven absolutely true. Before it's proof a fool might suggest that it appears true to all tested variables thus far, so it must be true. But mathematical proof requires it be demonstrably true for ALL variables. That is we cannot accept it as true even if it appears so, until it is shown true for every value.

[black magick hustla]

][/B]Marmot, whilst I find what you have said is interesting, I am concerned with the boundaries of science and its application. As such, treating all of science as a language game does not answer these questions. I am not too interested in what x group of people think about science and how they use the words particular to them....I am trying to establish logical limits on a scientific method, distinguish this method as it is from how it is conceived of, and illustrate problems. I guess I am talking about science as a 'philosopher of science' as would be evidenced by the questions asked. Consequently I am not asking about any particular view of science, I am aware of many of these, what I am looking for is a good desription of what science actually is. That is not according to a realist, or instrumentalist, or hypothetico-deductivist etc...I know what these people say on the matter. I am more concerned with a reasonable and accessible account which could underpin the higher degree of 'belief' we place in whatever it is that we call science.

the whole issue of science being part of a language game is the heart of your question though. i dont think you can give a specific parameter to it, in the same way you can give a specific parameter about beauty.

the limits of whatever we call science are probably the limits of any sort of logical discourse - language. we cannot talk about multiple universes or a cat being both dead and alive without falling in nonsense.

[Iowa656]

we cannot talk about multiple universes or a cat being both dead and alive without falling in nonsense.

But logically the cat is both alive and dead at the same time.

Quantum mechanics may appear strange to you, but that doesn't make it false.

Relativity shows how time can be stopped by travelling at the speed of light, just because that seems absurd why does that mean it's incorrect?

150 years ago people said it was "nonsense" to believe in evolution.

[Hegemonicretribution]

the whole issue of science being part of a language game is the heart of your question though. i dont think you can give a specific parameter to it, in the same way you can give a specific parameter about beauty.

the limits of whatever we call science are probably the limits of any sort of logical discourse - language. we cannot talk about multiple universes or a cat being both dead and alive without falling in nonsense.

First off I am waiting until I have the time to give an adequate response to the others, one is forthcoming though, even if it is after I hand an essay in. I answer Marmot here because my point is somewhat shorter.

Well as I said before, if what science is, is contingent upon the adoption of a particular language game, then I am concerned about what science is to a philosopher of science.

As there is not much for me to respond to on this (at least not now) I will offer some general thoughts on the use of language games as determining what science is.
OK, you place a constraint upon the limits of science in terms of language. This does not describe what the limits actually are, nor does it further our understanding of whatever 'scientific meathod' is.

If you can't give a specific parameter, then how do we distinguish science from non-science? Sometimes we want to use a good 'scientific' source, sometimes we claim a source is pseudo-scientific, or just plain bullshit. If the scientific method is only sensible according to an individual group's language game then you offer no means to assert good and bad methods.

On such a view, why is creationism not science? Why is evolutionary biology? OK, you might say that one group does not recognise the other as scientific, but I do not like the idea that external recognition is the distinguishing factor. This allows people to deny things they disagree with as unscientific, and claim those things they agree with are scientific. I get why you might appeal to language games to 'dissolve' this issue, but as I said it really does not get at what I am trying to establish. Would you tell a child that science is nothing more than the adoption of the appropriate language game? Would this help them actually understand what it is? It is the case that many people use scientific language here all the time, but if it is only according to their language game, then there is no means to decide between views. We have no means to say what constitutes a favourable method, and a poor one.

Slight re-phrasing, when people here talk about science, how do we decide between sources which reach different conclusions? Or do we accept that there is no means to because they are using a different understanding of science (one specific to the language game they are playing), and consequently the best explanation/predictive tool is a matter of perspective?

It seems that you implicitly say that science is like a broad pragmatism. Not a fixed approach, but simply a varried means of apprehending the external world. One which is not fixed by common actions, methods or approaches, but by a body of discourse which will use this label (science) differently depending upon which particular language group they fall into?? I do not know if this is what you are saying, but is does seem to follow from what I am understanding of you.

I will continue this in separate posts if you want, but I will say that for my responses to the others I am assuming that there is a 'scientific method' and it is one we can talk about meaningfully, or at least usefully.

[Die Neue Zeit]

Try looking up induction and constructivism to get more info on the limits of the scientific method.

[Hegemonicretribution]

Mikelepore

In science, The things about which we think and communicate are objective features of the world. But every symbolic notation to permit thinking about them and communicating about them depends on the culture and individual idiosyncracy.
This does not appear to answer the question. Unless you are saying they are objective, but our means of communicating aren't? But then again you think that thinking itself is somehow dependent upon this notation? If so it would not appear to be the objective feature of the world that may have been suggested. Also I do not accept that notation varries as much as you have suggested, at least not now where there are a few deviations from an otherwise accepted notation.


"More important" isn't a factual matter. It's a description of someone's state of mind. Perhaps an architect only cares about whether a building will stand up. There's nothing wrong with that either.The builder cares about building, the scientist cares about science....you specify a concern that the building should stand up, what is the concern of the active scientist? Perhaps 'more important' may be subjective to some extent, but I do not think it is entirely. I highlighted some of the most important aspects of science, I am asking for normative claim about these and their role in a useful scientific method. If they are not 'more important' then how is it you actually conceive of the relationship between science, explanation and prediction?


My point is that doesn't have to be any hypothesis at all for a task to be a useful part of the process of science....
I do not think we are in too much disagreement here, although the one I would claim is that even collecting raw data is contingent upon some assumptions at some level. I would also contest the notion that collecting 'evidence' irrespective of a hypothesis is as common place as it has been assumed. Perhaps your claim is that this is how science ought to function? Essentially though I can accept most of what you have said.


Any irregularity becomes just another regularity as soon as you determine its extent. If the measurement values jump around erratically, how much deviation is there? Did tweaking any parameters cause the mean or standard deviation to move? Now you have another regularity to report. But since I started out mentioning the taxonomy of animals, I'll stick with that. Irregularity appears in the form of an exception to an earlier block on the chart. First you may say that fish have bones. Then you correct the statement for a subclass of cartilaginous fish that don't have bones, which probably becomes the favorite specialization of some other researchers. We're not losing the pattern recognition that we have for understanding nature, but just filling out the picture.

This appears to assume a-priori thateverything is ordered. As for the fish your position changes from, 'all fish have bones,' to 'not all fish have bones.' You are adapting a framework to account for previous erroneous conclusions. This I do not mind, but I feel this categorisation gives a somewhat limitted picture of science. What you have here is an entirely descriptive account; it is only predictive insofar as past observations constitute a hypothesis. Even then it does not offer explanations for specific differences. Now I understand that these clasifications need to come about for a particular branch of science to develop, but that does not mean they are all there is. Consider my questions as analogous to 'what is the method of football.' I am not concerned with what a goal keeper does, or a striker, although I accept that these activities are a part of football. I am interested in the overall practice and how it is differentiated from, say rugby. Now in such a case you would just point to methodological differences, and use this to establish distinguising features. Now distinguishing science from social science is one of my concerns, but I am also interested in distinguishing it from poor scienctific method, historical approaches, everything that is not involved in an 'ideal' scientific approach. I accept this 'ideal' may be a fallacy, but I am only introducing it as a goal; I am looking for the best case scenario about science.




Neither has the manner in which science 'seeks to find out about' regularity. It is this with which I am more concerned. That's a procedural issue, which you can read about in an encyclopedia or something.

:lol: This is what the thread was supposed to be about. If science is granted superior status, then what is it about its procedure which makes it worthy of this status? You can ignore most of what I have written if you like, because I think we have been arguing about two different things. I am not going to look in an encyclopedia, because it does not give the kind of depth I was looking for here. I was not trying to learn about a particular procedural issue for scientists, I am trying to instigate debate about what this procedure ought to be. I am sorry about any lack of clarity.


Social sciences follow the scientific method to the extent that the tools are available. We wouldn't allow sociologists to keep human societies locked in glass containers for them to conduct controlled experiments on, so that tool isn't available. Identifying correlations is less revealing, but at least its a tool that is easily available.
I am sorry, but I have to disagree here. Social sciences follow a whole host of methods. Some think they should emulate natural sciences, others do not. Qualitative descriptions are often seen as more valuable than quantitative data because there is a feeling that humans may well be more than just a particular recorded result. Others deny this view and adopt a more positive approach to research. The very concept of triangulation of methods would stand against you here I think. Also, you cannot lock planets in a lab either, but we manage to talk about them fairly successfully. That said many social sciences do lock people in labs, but it is said taht such conditions do not allow for a representative observation, because the 'controlled environment' is itself unnatural to the human subject. In others words because humans do not exist or function in a controlled environment, it would be unjustified to treat results of observations of humans in such an environment as weighty. There is much to say on this, but I fear it may take us to far afield.


His conclusion had to be that atoms are not filled with stuff throughout their volume, as the old "plum pudding" model of the atom had supposed, but, rather, the atom is mostly empty space, with most of its mass contained within a small nucleus. Furthermore, it had to be that it's the positive charge in the atom that resides in that nucleus, and not the negative charge, to account for the repulsive deflections.
It did not 'have to be' at all. Other explanations may yet come to light; I am not sure whether such a trascendental deduction counts as sound scientific method. People have used this to justify god before; look at creationism.

That said I think that it was good scientific method overall, I just do not accept that it ends there. The refutation of the plum pudding model is far stronger that the 'proof' of the new model. Certainly aspects of it are very appealing, but as a complete and literal description further tests would be required.


By the peculiarity of grammar, we may phrase it to say that he disproved one idea, or we may phrase it to say that he proved a different idea.
He disproved one idea, and gave a very convincing case for another. I do not see where the proof comes in, unless you use a somewhat weaker sense of the word. This is a very nice case to consider, but unfortunately most science does not function like this, although it is good when it does. Often the old account is simply too inadequate to be deemed sufficient, and a new account is presented as a solution. It does not mean that the new account will be readily accepted. Indeed if it is in its early stages it too may face problems. There is a tendency in the history of science to emphasise cases such as the one above as a success story. The perihelion of mercury, or the discovery of Neptune are two other cases which are often over-stated as typical to the scientific method. Now I am not necessarily denying these are ideal cases, but I am denying that much of what passes for science resembles this state of affairs.

More generally the difference between proof and disproof is not simply two sides of the same coin. The difficulty regarding proof is best illustrated through problems with induction. Disprooving something comes somewhat easier.....unless of course allow for a certain understanding of auxiliary hypotheses and view the identity of the hypothesis in a certain way. Nevertheless, a degree of disproof comes much more readily than a degree of proof.

Next reply coming soon..

[Hegemonicretribution]

That is indeed correct, it's a lot easier to disprove a statement that it is to prove one. Whatever my statement is, one piece of evidence can disprove it, but an "infinite" amount of evidence is needed for definite proof.
Well not necessarily infinite, but every piece of evidence (this may itself be infinite)....I take your point though, induction does cause a problem. I am looking for scientific approaches in the light of the challenge of induction. What degree of evidence would you assume one observation to provide for example? Even if you cannot achieve certainty, surely you can increase you belief in proportion to the evidence? If so how would you suggest this?

Mathematical proofs may be somewhat easier to come by, it is a shame science cannot always rely on similar proofs.


That depends on how your mind works.
Are there objective factors we could appeal to?


In many other conditions however a less conclusive proof can be accepted.
Is this considered a proof or merely a best explanation?


If you are a judge is 99.99% chance sufficient to convict me? If it is then one in 10,000 cases are statistically likely to be a false conviction.
It is the degree of belief that we could reasonably place in such claims that is of primary concern, because this is more typical to science than mathematical proofs.


Hypothesis z is a extension of hypothesis y, usually including hypothesis y in it.
What if it doesn't? Or if it is a modification, what if it has been a trimming down of the hypothesis rather than a building up?


For example, my first hypothesis, hypothesis y, is that all matter is made out of atoms. I conduct many experiments and they present the fact this is true. I can mix chemicals and produce different elements. I can break down matter into atoms. I then accept the hypothesis.

In each case the assumption first and foremost is that there are atoms. What if I said that all matter was made of otoms? The same experiments could presumably support this also? It is the invocation of the theoretical entity which brought about the experiments and evidence to add support to the ideas concerning the theoretical entity. Circular perhaps?


Next, and in fact hundreds of years latter, hypothesis z is that atoms are made of hadrons and electrons. I do experiments in a particle accelerators, using radiation or using charged particles that demonstrate this to be true. I then accept this hypothesis z. Does this mean that hypothesis y is false? Not really; it is true, yet is insufficient.

That is not so problematic, but as I said before what if Hz was lacking in certain flaws of Hy (instead of simply containing a few modified ideas)? So for example Hy was concerning a plum-pudding model of the atom, and Hz the modern conception. Whilst there may be some cross over between y and z, it could not be said that y is simply contained within z.


Evidence isn't really produced, more accurately would be a description that is it observed. Example, I drop a ball, I observe it, it falls to the ground, I repeat, that is my experiment. I have produced data; all objects fall to earth.


This is not as troublesome as I thought before. I think I misunderstood what you meant by produce.



Oh indeed, no one would claim X does this, so everything does this.


I did not mean simply that you need more than one observation, but I was highlighting a problem to do with invoking a theoretical entity which then becomes part of your observations. Not so problematic in itself, but when you assess competing theories, it may be that they do not invoke the same theoretical entity, and so comparison becomes very difficult if not impossible.



Usually in science it is the other way round. The evidence is this, so the theory is built upon it. In the flat earth example, the evidence was that earth seemed flat to our perceptions and that ships rarely returned form long journeys, thus the theory was that the earth was flat. New evidence became apparent and flaws were found in this first theory, and such it was improved.

Science makes use of pre-theoretical data, yes, but I am not sure how much this compares to post-theoretical observations. Also there are a lot of takes on science which would deny that you collect data and then build the theory. Hypothetico-deductivism (as probably the most famous) would suggest that you errect the theory, whatever it is, and then try to disproove it. Theory first. The point is that whatever problems arose out induction originally will only be realised again at the level of theory generation. Your considerations at that time would be those of a time before you can explain what you intend to explain. In other words you may well come up with completely shit theories, and waste time trying to proove them.


One counter example is enough to debunk the whole theory.
Just to remain as devil's advocate here....is it though? One counter example may just require an auxiliary hypothesis no? Classic example, plotting the orbit or Neptune.

I am looking for Fermat's theorum now, thanks for the tip.


But logically the cat is both alive and dead at the same time.
You can argue that it is. I do not know if there is a prima facie case for this though. Perhaps there is some proof of it of which I am not aware, I am not sure that it is as intuitively true as you have suggested.



Quantum mechanics may appear strange to you, but that doesn't make it false.

True


Relativity shows how time can be stopped by travelling at the speed of light, just because that seems absurd why does that mean it's incorrect?
I am not sure I agree with this....but I do not think that it disprooves relativity.



150 years ago people said it was "nonsense" to believe in evolution.
Do you think they will again in 150 years? ;)


JR:

Try looking up induction and constructivism to get more info on the limits of the scientific method.

Anything in particular? I am fairly familiar with both, but if you have any particuarly interesting sources that would be good.

[Iowa656]

Well not necessarily infinite, but every piece of evidence (this may itself be infinite)....I take your point though, induction does cause a problem. I am looking for scientific approaches in the light of the challenge of induction. What degree of evidence would you assume one observation to provide for example?

One observation has little accountability. There are many issues with generalising results, as I am sure you are aware. If a result is continually observed then it is generally accepted to be accurate. If you repeat an experiment 100 times and the results are the same then I'd argue that the results should be accepted.



Even if you cannot achieve certainty, surely you can increase you belief in proportion to the evidence? If so how would you suggest this?

Certainty is something that doesn't really exist. The basic answer though is through repetition, if an experiment is repeated many times in different circumstances and the results are what is expected then the theory should be held as true.



Mathematical proofs may be somewhat easier to come by, it is a shame science cannot always rely on similar proofs.

That's the fundamental difference between Mathematics and certain aspects of science. Pythagoras' theorem will be as true in 200 years as it was when it was first wrote, our understanding of the Big Bang will be vastly different even in 20 years time. That's not to say mathematics is static and rigid, because there are thousands of new theories each year, but when a new mathematical theory arises, if the proof is up to the standards, it's accepted generally without question. A new theory in biology would certainly not hold such a position.



Are there objective factors we could appeal to?

I don't think there are. How do you prove something? If it is taken that something must be proven 100% before accepted then you ought to reject a lot of scientific knowledge you have.

Similarly if a new drug increases survival rate from a certain disease in a sample from 35% to 40%, does that prove the drug is effective? If I toss a coin many times and get 60% heads, does that prove the coin is weighted? To answer these questions you have to ask one fundamental question, what is the probability of getting these results from random chance alone? Then you have to set an arbitrary limiting value, i.e. if it is shown that there's only a 5% chance of these results occurring by random can we accept them as significant? Is it more accurate to use 1% chance? 0.001%? Therein lies the problem. If you can't definitively prove a postulate true for all cases then your acceptance of it is based merely on chance.



Is this considered a proof or merely a best explanation?

That depends entirely on the specific case and what it presents. A theory is usually the best explanation for the observations that can be made. When does the best explanation become the explanation? If my theory accurately predicts outcomes time and time again is it then true?



It is the degree of belief that we could reasonably place in such claims that is of primary concern, because this is more typical to science than mathematical proofs.

Exactly, and frankly there's not one answer.



What if it doesn't? Or if it is a modification, what if it has been a trimming down of the hypothesis rather than a building up?

The great scientific revolutionary papers have usually been ones that put forward an entirely new position, one's that asked questions that had never been asked before.

Having said that nearly all scientific theory's though are an evolution of some sort of previous theories. The new theory may build upon, criticise, refute, or even turn sideways the old theory.




In each case the assumption first and foremost is that there are atoms. What if I said that all matter was made of otoms? The same experiments could presumably support this also? It is the invocation of the theoretical entity which brought about the experiments and evidence to add support to the ideas concerning the theoretical entity. Circular perhaps?

"Otoms"? It's best not to question the entirely arbitrary names scientists have attributed to objects or phenomena. "A rose by any other name would smell as sweet". No not entirely, one idea in the early 1900s was that there was this "ether" that existed trough which light travelled. That was the theory. It was put to the test in the ground breaking Michelson–Morley experiment. The experiment showed that the "ether" did not exist. The theory was disproved, shown to be incorrect. Countless times this has happened. To suggests that testing a presented theory must lead to it's acceptance shows a great misunderstanding.



That is not so problematic, but as I said before what if Hz was lacking in certain flaws of Hy (instead of simply containing a few modified ideas)? So for example Hy was concerning a plum-pudding model of the atom, and Hz the modern conception. Whilst there may be some cross over between y and z, it could not be said that y is simply contained within z.

The plum pudding model did in fact suggest that firstly matter was made of atoms and atoms inside them has small electrons. The modern model still presents this as factual. There where radical changes in the understanding of the atom, but everything previously thought was not rejected.



Apologies if my wording is causing confusion.



I did not mean simply that you need more than one observation, but I was highlighting a problem to do with invoking a theoretical entity which then becomes part of your observations. Not so problematic in itself, but when you assess competing theories, it may be that they do not invoke the same theoretical entity, and so comparison becomes very difficult if not impossible.

Oh indeed this is true. Many competing theories of the String theory suggest that there somewhere between 11 and 23 dimensions. I challenge anyone using present knowledge to determine which, if any, are true. They won't be able to. What does that mean? Mainly that present technology is not advance enough to determine the validity of any opposing theory.




Science makes use of pre-theoretical data, yes, but I am not sure how much this compares to post-theoretical observations. Also there are a lot of takes on science which would deny that you collect data and then build the theory. Hypothetico-deductivism (as probably the most famous) would suggest that you errect the theory, whatever it is, and then try to disproove it. Theory first. The point is that whatever problems arose out induction originally will only be realised again at the level of theory generation. Your considerations at that time would be those of a time before you can explain what you intend to explain. In other words you may well come up with completely shit theories, and waste time trying to proove them.

Both techniques are used. But it seems unlikely someone would theorise on a subject they have no previous association with or understanding of. Netwon never questioned how time alters in a gravitational field because he never observed it's effects. Neutrino's where not discovered for years because they are barely visible. But can you accept a theory that cannot be disproved? For example, my theory is that there is a completely un-observable entity in the universe that's only purpose is to observe the universe. Even even how about God. Impossible to disprove, but does that prove it? That is defiantly true, many, even great, scientists have spent a long time attempting to prove things we now accept as false.



Just to remain as devil's advocate here....is it though? One counter example may just require an auxiliary hypothesis no? Classic example, plotting the orbit or Neptune.

Yes it is. For example, imagine next week scientists found a material that had had mass but did not obey the laws of Gravity. Everything we ever knew about gravity would instantly become worthless, because the very first assumption is that all mass is effected by gravity. If an observation does not align with the theory it is the theory, or some flawed process in the act of observing, that is incorrect, not the universe. Present theories can not explain many things we can observe, the biggest example being the lack of anti-matter in the universe, unfortunately that means the theory is incomplete. There may very well be presently unexplained problems in Neptune's orbit, although I'm not sure that's even true, the most probable explanation would be that there are objects we have not yet observed.





You can argue that it is. I do not know if there is a prima facie case for this though. Perhaps there is some proof of it of which I am not aware, I am not sure that it is as intuitively true as you have suggested.

In all honestly my understanding of Quantum Mechanics is somewhat lacking. There may be a solution to the somewhat paradoxical situation.




I am not sure I agree with this....but I do not think that it disprooves relativity.

What part are you disagreeing with? Relativity has been tested and is continually shown to hold true. Don't forget that it is in fact impossible to travel at the speed of light, you would need infinite energy, something that obviously is not possible. So this situation would never occur.




Do you think they will again in 150 years? ;)

Honestly I don't. Without going into specifics evolution has a lot of evidence backing it, with more evidence becoming available all the time. Yes there are current problems with the theory I admit that, but these are merely a lack of finding the evidence that is so well buried. In 150 years evolution won't be exactly what we think it to be presently, but it will built upon the same basics. In the same way Einstein updated Newton, I feel it inevitable a future biologist will update Darwin. Einstein didn't say Newton was wrong, just that he missed sections.

[Hegemonicretribution]

FUCK FUCK FUCK FUCK FUCK!!!!!!!!!

Sorry, I just replied to every line, and then I closed the browser before I hit post and lost everything. I would like to continue this discussion because I feel we were just getting to the interesting questions, and agreeing on issue as well, but I have not the heart to do this right now having wasted hours today.

Quick question, are you familiar with Bayes theorem? It is just because it will play some part in my eventual answer. Apologies for being so long about this.

[Invariance]

My definition is: Science is recognition that there is some regularity in the universe, and the efforts to determine under what conditions the common patterns reappear. If it's possible to experiment with something, you do. If you can't do experiments, at least you try to name the classes of things and list their properties.


Such facts are regular patterns in the world. This regularity is the subject of science. Science seeks to find out about it.

I don't think there needs to be any regularity in science, or for it to identify patterns, even though many sciences very well may. Your conception seems to favour the more physical sciences which are more 'law orientated' versus, for example, biology, which some have argued has a more 'historical' approach.

If biology posits that x species will evolve characteristic y because of z, to what extent can such an event reappear (to use your phrase)? Clearly, repeatability is very limited in such a circumstance, unless we take a very broad definition of repeatability, which would have less empirical worth. To what extent can the Alvarez hypothesis claims about facts be repeated - that dinosaurs would be wiped out by by the impact of an asteroid and species x, y and z would be dominant because of that? But would you be prepared to say that they aren't doing science because they aren't looking for a pattern? I know I wouldn't be.

[Invariance]

I am looking for scientific approaches in the light of the challenge of induction. What degree of evidence would you assume one observation to provide for example? Even if you cannot achieve certainty, surely you can increase you belief in proportion to the evidence? If so how would you suggest this?

Firstly, this is a huge question, and I won't do it justice.

There have been a number of approaches to this problem.

The first is to accept induction as illegitimate. Hume is a good example of this approach and Popper also, as I will explain below. I think instead of trying to solve it, you could accept the problem and not avoid it; 'I don't reject inductive reasoning, I just recognise its necessary limitations.' But I also agree with Quine that the whole analytic/synthetic distinction isn't very useful (read Two Dogmas of Empiricism). Rejecting inductive reasoning may often seem unreasonable, but rejecting counter-inductive reasoning as reasonable. Yet that is odd.

Let me explain. Where induction assumes that the future will be like the past (for example, the past instances of the sun rising surport the fact that the sun will rise tommorrow), counter induction predicts that the future will be unlike the past. If a gambler has lost 15 hands in a row, he may infer from that past results, that they should play the next had since they are 'due' for a good hand!

Any normal person would probably give them an odd look; past results have no bearing on a present hand. Flipping ten heads on a coin is no evidence that the next flip will be a tails (or a head).

If we think accepting counter inductive arguments is wrong, why should we accept inductive arguments as right?

To borrow Russell's case, if everyday a farmer rings a bell and then proceeds to feed the chickens, and this goes on for a number of years, then one day the farmer rings the bell and cuts the chickens head off, the chicken has made a fatal mistake in causally linking the ringing of the bell to being fed, and not getting its head cut off. If the chicken shouldn't trust induction, then why should we?

There was another approach, by Popper, which essentially accepted Hume's argument but rejected its importance. Remember, to Popper, confirmation wasn't important, only falsification - one black swan falsifies the statement that all swans are white. And as Carl Hempel (in my mind humorously) pointed out, a blue fish is also evidence that all ravens are black.

P1: All ravens are black.
P2: Everything that is not black, is not a raven. (Think about it).
P3: This is a blue fish (and hence, not a raven!).

Since statement two is logically equivalent to statement one, statement three does in fact count as evidence that all ravens are black (even if it is an extremely minute amount of evidence). This may seem absurd, but that's only because of our mistaken psychological illusion that the statement “All ravens are black” is only about ravens; it isn't.

Yet note this: since Popper accepted Hume's argument, the very best that a theory can do is survive attempts to falsify it. But surviving attempts to falsify does not make a theory true or confirmed. Popper called this corroboration. Moreover, Popper had to deny that a theory's corroboration indicates its future success since that is also a form of induction!

But why should we accept corroborated theories over uncorroborated theories? The mere fact that a theory has not been attempted to be falsified does not mean it is wrong, it just means it hasn't been tested. Indeed, for some experiments it took several hundred years before they could be falsified. Popper thought scientific theories aim at the truth but never get any evidence that they have attained the truth. Of course... science has never worked like Popper proposes, or if it has only in moments of crises, but since you asked...

There was another 'solution' to the problem of induction. Perhaps you have heard of the law of large numbers, which essentially (non-mathematically) states that by "taking a large enough random sample of a population, we can attain as high a probability as we would like of coming as close as we would like to knowing the frequency of a trait in the population."

A random sample is defined as being one which has the same likelihood as being picked as any other. There are a number of problems with this. Firstly, trying to obtain a random sample. For example, if we were to randomly select a number of telephone numbers from the book (using whatever mathematical method of randomness you choose), and proceed to ring those people at midday and obtain whatever information you want, to what extent can you call that random? There are problems, of course - you have necessarily restricted your data to those with telephones. You have also restricted your data to those who are at home at that particular time. So you've already essentially rejected any homeless person, as well as restricted it to people who were at home at midday - which would bear all the distortions of that - for example, you may obtain a higher number of housewives versus men which would distort the data.

To give another example, when we are trying to make a claim that all copper conducts electricity (the standard example philosophers use), we are making not just a claim about copper on Earth today. But copper everywhere and for all time. How arrogant can we be to claim that our minute planet, and our minute time can constitute a sufficient sample of randomness! On the contrary, it looks totally non-random in that consideration...

I'll stop there. There are other areas of discussion, like the hypothetico-deductive model, the inference to the best explanation and what are known as Mill's Methods. They all have their problems, and science certainly shouldn't be constricted to them. It also begs the question if science should be constricted by any method, and if I get a chance I'll take about Feyerabend's approach later, which I don't think you shouldn't be so quick to dismiss.

[mikelepore]

I didn't say that events are repeated -- I said that common patterns reappear.

In the case of evolution, the pattern isn't that species x would necessarily evolve characteristic y. The pattern is that when characteristics in general have evolved we so often find that those characteristics were the type that would have increased the probability of reproducing. Whether this was done by changing the means of finding food, attracting mates, eluding predators, making tools, etc. is random because mutations are random. The fact that most mutations are harmful and those affected individuals more likely die is also an important conclusion.

In the case of the dinosaurs, the patterns would be how a reduced rate of photosynthesis affects the rest of the food chain, and how omnivores and detrivores had greater rates of survival than herbivores and carnivores, etc. The exact outcomes were random. To make a comparison between the Cretaceous-Tertiary extinction and the Permian-Triassic extinction is also a search for commonalities.

Also we don't have only aspects of biology that look backwards at a probabilistic history -- evolution and extinctions. The storage of energy in ATP, or the Krebs cycle, or the endocrine system, are chemical processes.