# Induction is Impossible Induction has many flaws, two of which are: 1. [Prediction is not Explanation](Prediction%20is%20not%20Explanation.md) 2. [Russells Chicken](Russells%20Chicken.md) While flaws (1) and (2) illustrate the fact that repeated observations cannot [*justify*](Justification.md) theories, it still lets inductivism off too lightly. For it still accepts the basic misconception that inductive extrapolation of observations to *form* new theories is even possible. It is not. It is impossible to extrapolate observations unless one has already placed them within an explanatory framework. [All Observations are Theory Laden](All%20Observations%20are%20Theory%20Laden.md). The same set of observations can be extrapolated in vastly different ways depending on the explanatory framework they are placed within. For [Russells Chicken](Russells%20Chicken.md) to 'induce' its false prediction, it must have first had in mind a false explanation of the farmers behavior. Perhaps it guessed that the farmer harbored benevolent feelings towards chickens. Had it guessed a different explanation - that the farmer was trying to fatten the chickens up for slaughter - it was have 'extrapolated' the farmers future behavior differently. Observations cannot be made without some sort of "context" associated with them, where that context is a "theory". The way in which the new set of observations will be extrapolated to predict the farmers future behavior depends entirely on how one *explains* it. By the benevolent-farmer theory, it is evidence that the chickens have even less to worry about than before. But by the fattening-up theory, the behavior is ominous - it is evidence that the slaughter is imminent. The fact that the same observational evidence can be ‘extrapolated’ to give two diametrically opposite predictions according to which explanation one adopts, and cannot justify either of them, is not some accidental limitation of the farmyard environment: it is true of all observational evidence under all circumstances. Observations could not possibly play either of the roles assigned to them in the inductivist scheme, even in respect of mere predictions, let alone genuine explanatory theories ### Counterargument But wait, one may say: "I constantly perform induction in my day to day life! For example, I see that traffic has been terrible on my commute home from work for the past 10 days. Tomorrow I predict, via extrapolation, that the trend will continue. I therefore take a different route". Surely that is an inductive approach? Yet a closer examination shows that even here induction was never truly being used. For this observation and the associated prediction were both *theory laden*. In order to predict that the traffic will continue, you must have been making use of implicit theories: that during the past ten days there had not been construction occurring that will have finally just ended; that no new road was installed over night; that there wasn't some conference in town that had been causing the traffic, which just wrapped up; and so on. It was in the context of these theories that you extrapolate the traffic will continue tomorrow. If you had an additional theory stating that tomorrow the construction that had been causing the traffic will have ended, you may predict that travel will be more smooth and fluid than ever. The key idea here is that you did *not* form a new theory based purely on inductive extrapolation. You formed it based on your best theories. These observations can of course fit in and play a role. But on there own they tell us nothing. --- Date: 20241011 Links to: [Science Does Not Work Via Induction](Science%20Does%20Not%20Work%20Via%20Induction.md) [Induction](Induction.md) [Fabric of Reality](Fabric%20of%20Reality.md) Tags: References: * []()