While studying I read about the principle of underdetermination of scientific theories. I made some researches online but I am more confused than before. I read about the Quine-Duhem holistic thesis, but I do not know if is the same as underdetermination.
The essential issue is stated by W. Newton-Smith :
Can there be theories which are underdetermined by all actual and possible observations ? That is, can there be logically incompatible but empirically equivalalent theories ? (W. Newton-Smith and Steven Lukes, 'The Underdetermination of Theory by Data', Proceedings of the Aristotelian Society, Supplementary Volumes, Vol. 52 (1978), pp. 71-91+93-107 : 71.)
That's okay but too brief. A further citation may help :
It is obviously true that at any given stage of a scientific enquiry the available data will in principle be compatible with many different, mutually incompatible theories. This is because theories always outstrip the data on which they are based, if only by universal generalization - the inference from data to theory is always deductively invalid. This point is sometimes expressed by saying that scientific theories are inductively underdetermined by the data.
Inductive underdetermination is not what most philosophers of science have in mind when they discuss the underdetermination of theory by data. In recent discussions, 'underdetermination' usually refers to the idea that there may be theories between which no possible evidence can decide, not merely no actual evidence. If two theories are underdetermined in this stronger sense, then however much empirical data we collect in the future, we shall never be able to decide between them on empirical grounds. I use the term 'strong underdetermination' to refer to situations of this sort. Where I use the term 'underdetermination' without qualification, it refers to strong underdetermination, not inductive underdetermination.
Why should it be thought that scientific theories are typically, or indeed ever, strongly underdetermined by data? Many philosophers believe this because they think that for any scientific theory there always exists an alternative empirically equivalent rival theory. Empirically equivalent theories are those whose empirical or testable implications are identical. Some authors treat the concepts of empirical equivalence and underdetermination as interchangeable, but I do not follow their lead. If two theories T1 and T2 are incompatible but empirically equivalent, I see that as a possible reason for thinking them strongly underdetermined; but the former state of affairs is not identical with the latter. The rationale for driving a wedge between 'T1 and T2 are empirically equivalent' and 'No possible evidence can decide between T1 and T2' will become apparent. (Samir Okasha, 'Underdetermination, Holism and the Theory/Data Distinction', The Philosophical Quarterly (1950-), Vol. 52, No. 208 (Jul., 2002), pp. 303-319 : 304-5.)
I'll refer for convenience only to Quine :
** Quine's "Two Dogmas of Empiricism" is so well-known that we may safely presume some familiarity with QE-holism, the informational holism that Early-Quine sets out in this essay. So let us begin our examination of Quine's holisms by enumerating, in Early Quine's own words, the most salient features of QE-holism:
"The dogma of reductionism survives in the supposition that each statement, taken in isolation from its fellows, can admit of confirmation or infirmation at all" ("Two Dogmas of Empiricism," p. 41).
"Our statements about the external world face the tribunal of sense experience not individually but only as a corporate body" ("Two Dogmas of Empiricism," p. 41).
"Total science is like a field of force whose boundary conditions are experience. A conflict with experience at the periphery occasions readjustments in the interior of the field. . . . But the total field is so underdetermined by its boundary conditions, experience, that there is much latitude of choice as to what statements to reevaluate in the light of any single contrary experience" ("Two Dogmas of Empiricism," pp. 42-43).
"Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system." ("Two Dogmas of Empiricism," p. 4)
(Gerald J. Massey, 'QUINE AND DUHEM ON HOLISTIC HYPOTHESIS TESTING', American Philosophical Quarterly, Vol. 48, No. 3, W. V. Quine Centennial (JULY 2011), pp. 239-266 : 253.)
The Quine-Duhem thesis comes into play when, for example, a hypothesis yields an observational result which conflicts with what the hypothesis predicted. Quine's point is that since it is never a single observation or set of observations which is involved but a set of assumptions on which the hypothesis rests - "Our statements about the external world face the tribunal of sense experience not individually but only as a corporate body" - a hypothesis can never been defeated by a single failed observation or any set of such observations since it may always be possible that it is not the hypothesis that is at fault but something in the 'corporate body' (theoretical assumptions, auxiliary hypotheses and so on) of which the hypothesis is a part.
So while the uderdeteration theory rests on the claim (a) that at any given stage of a scientific enquiry the available data will in principle be compatible with many different, mutually incompatible theories, the Quine-Duhem thesis is in a different line of business, (b) concerned with the defensibility of a hypothesis when the observations it predicts fail. Adjustments can always be made in the 'corporate body' to explain the failure and preserve the hypothesis.
Hope this is of basic help.
Boiling down the SEP entry, this is what I get.
Underdetermination is the case where we just don't know enough to get to a conclusion. Consider the link between smoking and lung cancer. We have concluded that smoking can cause lung cancer, but if we just looked at the correlations, we could come up with other explanations. Perhaps a gene or combination thereof makes people find tobacco more addictive and lung cancer more likely. Perhaps an incipient cancer creates a subtle discomfort that can be alleviated by smoking. We know more than just the correlation, and therefore having more information allows us to know that smoking causes cancer.
The entry then divides underdetermination into holistic and contrastive varieties.
Holistic underdetermination, in science, considers that it's normally impossible to test scientific hypotheses by themselves, but only in conjunction with other hypotheses. Obviously, interpreting certain things that happened at the CERN accelerator, by themselves, as evidence of the Higgs boson required a large mass of other theory, which could be wrong somewhere.
The example in the entry is orbital mechanics. In the early 1800s, there were theories of orbital mechanics based on Newton's laws, and there was our knowledge of the planets in the solar system. Both the orbits of Mercury and Uranus had anomalies. In the case of Uranus, our theory was correct and we didn't know about Neptune. We found it in 1846. Similarly, Mercury's anomalies could be accounted for by a planet closer to the Sun, named Vulcan. Nobody was able to observe Vulcan, however, and eventually we found that it didn't exist as predicted, but the theory was wrong. The anomalies were accounted for with a new system of physical laws, relativity, that differed from Newton's in some ways.
Therefore, whether our theories were wrong or our observations incomplete couldn't be determined by information available at the time, and in this case the same sort of anomalies were caused by entirely different things.
Distractive underdetermination questions whether we have all the possible hypotheses. It's conceivable that we would eventually have found that Neptune didn't exist and relativity was wrong, and that the anomalies were caused by supertechnological Martians playing practical jokes. What other hypotheses are possible? Suppose we had two hypotheses that predicted different things for different experiments. We still can't say that one is right, because there may be other hypotheses that fit the facts which haven't occurred to us.
Light was clearly made of discrete particles or continuous waves. Early physicists ran experiments and found it to be made of waves. Therefore, it wasn't made of particles. Then the anomalies started popping up.
When something is heated up enough, it starts to glow, and the color of the glow depends on the temperature. Blacksmiths used the glow of iron as something of a thermometer, although they referred to temperatures like "dull red" and "cherry red" rather than anything quantitative. It turned out that, if life was continuous waves, there was no way any physicist could explain how that worked. Max Planck showed that, if light came in discrete units, with energy going up as frequency increases, then a black body would only be able to radiate light up to a certain frequency, and the calculations worked.
Similarly, when light hits certain materials it can cause them to emit electrons of certain energies. The number of electrons varied with how bright the light was, but brighter light didn't produce more energetic photons. Einstein found that, if light came in packets like Planck supposed, that this photoelectric effect could be understood.
Eventually, physicists realized that light was, in a highly non-intuitive sense, both waves and particles, and properties of either would dominate depending on exactly what people were doing with it. This third hypothesis is what we're using today, and it seems to be very successful.