Imre Lakatos: Scientific Research Programs

Lakatos' "Core" and "Periphery" Model of Scientific Theories
Progressive/Regressive Problem Shift and Theory Replacement Questions for Study

Imre Lakatos (1922-1974) was born and lived in Hungary until he left in 1956 in the aftermath of the Soviet military intervention in that country. He was an anti-fascist partisan during the Second World War, and thereafter, rose to a high position in the communist government which ruled the country after the war. However, he was dismissed from office and imprisoned in the early 1950s. He sided with the democratic forces in the mid 1950s, and became an exile in England after the repression of the freedom movement.

Lakatos' "Core" and "Periphery" Model of Scientific Theories

Lakatos' theory of scientific change is not unrelated to his political experiences, though it stands, theoretically, on its own. Lakatos distinguished between two parts of a scientific theory: its "hard core" which contains its basic assumptions (or axioms, when set out formally and explicitly), and its "protective belt", a surrounding defensive set of "ad hoc" (produced for the occasion) hypotheses.

For example, we have seen that the "core" of Aristotelian cosmology was constituted by the basic propositions that (i) the earth is motionless at the center; (ii) the sun moves around the earth; (iii) all motions of the sun, planet and stars are circular.

But the existence of the retrograde motion, variable distances and variable velocities of the planets were, in Kuhnian terms "anomalies" which should have opened up a "crisis". In Popperian terms, these should constitute "refutations". But in Lakatos' terms this crisis or refutation was forestalled for nearly 2000 years by the adoption of a series of "protective" measures, which we have already discussed: epicycles, eccentrics, and equants.

Logically speaking we have the following, with

In Popperian terms: If A then B; But not-B is the case; therefore not-A. In words: the core assumptions (A) imply "normal" observations (B); but there are anomalous observations (not-B); therefore, the core assumptions are wrong (Not-A). In other words:

The core Aristotelian assumptions Imply
No Retrograde Motion, No Variable Distances and No Variable Velocity.
But these do in fact occur, and are regularly and reliably observed.
Therefore The core Aristotelian assumptions are wrong. (by modus tollens)

If this were the case, the core assumptions would have to be dropped immediately, as Popper states, or a crisis would open leading to a new paradigm, as Kuhn claims. But this did not historically happen, so Lakatos claims that the Popperian and Kuhnian models are not accurate.

In Lakatos' model, we have to explicitly take into account the "ad hoc hypotheses" which serve as the protective belt. The protective belt serves to deflect "refuting" propositions from the core assumptions, since the conjunction of the core with the periphery actually implies the otherwise anomalous statements. So, let's explicitly add in the ad hoc or protective hypotheses for cosmology, which are:

According to Lakatos, what happens, in logical terms, is the following: If A and C then B; And B is the case; So, A along with C is confirmed. The core assumptions and the peripheral hypotheses together imply what is in fact observed. In other words:

(Basic assumptions at the Core): Aristotelian assumptions Plus
(Ad Hoc hypotheses in Periphery): Epicycle/Eccentric/Equant Imply
(Observations): Retrograde Motion, Variable Distances and Variable Velocity.
Therefore, The core Aristotelian assumptions, aided and protected by the ad-hoc hypotheses, is confirmed.

Note that we cannot conclude from this that A and C are therefore true (this would be the fallacy of affirming the consequent), but we can say that A and C are probably true, or confirmed to some degree. (see the discussion in the section on Popper about confirming, rather than proving a scientific theory). The core assumptions A survive, albeit with the needed addition of the protective belt.

We can conclude as follows:

Progessive/Regressive Problem Shift and Theory Replacement

Lakatos adds an important caveat to his "core/periphery" model: The theory (core) must continue to make original predictions which are in fact confirmed. This is termed a "progressive problem shift" (in other words, the theory is progressive).

But if all that happens is that ad hoc hypotheses are continually being required to prevent new observations from condemning the theory, without the theory making new and true predictions on its own, then the theory is said to be in a "regressive problem shift" (the theory is regressive).

In the latter case, the theory is ripe for replacement by some other theory which has developed independently of it. For example, eventually the Copernican core (with Kepler's elliptic orbits as part of it) replaced the Aristotelian core, with its protective belt of mathematical hypotheses which had become cumbersome, inelegant, and regressive. Similarly, Einstein's theory of special relativity replaced Newtonian mechanics, which had failed to continue to progress and had become burdened with a large protective belt (needed to explain away problems such as the null result of the Michelson-Morley experiment).

Questions for Study

(1) What does Lakatos mean by "core"?

(2) What is the role of the "periphery"?

(3) How does Lakatos' theory differ from that of Kuhn?

(4) How does Lakatos' theory differ from that of Popper?