The New Riddle of Induction

American philosopher Nelson Goodman in response to Hume’s riddle of induction, created his own version of the problem, naming – The New Riddle of Induction. The main point of the philosopher was to show how flawed and contradictory can inductive reasoning be. In his thought experiment, Goodman presented a hypothetical substance called grue, which is anything that has been observed before a certain time in the future (for instance in November 13, 2050) and is green or is blue after the year of 2050 . If we say that all emeralds observed are green and the inductive reasoning lets us to conclude that all emeralds will remain green, hence they must also be grue, as it is still the time for emeralds to be green, which is part of the definition of grue. However, by confirming the fact that they are grue, we are simultaneously accepting the fact of emeralds being blue. It may seem that we are in a dead-end situation but many philosophers offered their solution to the problem, even though a decent answer to the problem has not been given yet.

Goodman distinguished two types of predicated — lawlike and non-lawlike. While making generalisation lawlike predicated are more likely to confirm the likelihood of valid conclusion, as they are based on some general principle. For instance, if we say this ice cream is cold hence all ice creams are cold, it seems to be legitimate, as coldness refers to an ice cream. The same structure doesn’t work for non-lawlike predicates, as the premise of “three people observed are eating ice cream, hence the forth one must also be eating an ice cream”. There is no general rule for confirming what will happen next based on our observation. What Goodman wanted, was to stress on the argument that induction works only when predicates form lawlike generalisation, but as we cannot tell which predicates will actually form lawlike generalisation, hence we cannot confirm whether induction work or not.

 The initial question that David Hume asked was can induction help us to know what exactly will happen in the future. By examining several cases, we make generalisations and assume that the same pattern will continue in the future. If the premise “all emeralds observed are green” concludes that “all emeralds are green”, which itself seems to be rather logical, but the conclusion of all emeralds being grue seems to be quite wrong. Just like we cannot be sure of emeralds not changing their colour and remaining green in the future, at the same time the assertion of emeralds actually changing their colour to blue cannot be denied either.

Goodman’s explanation for the the problem is entrenchment, according to with we are used to use green as a predicate but the same doesn’t apply for grue. In response to Goodman, British philosopher Swinburne offered his solution to the problem by stating that there are two times of predicates -- qualitative and locational. While having knowledge of something being green we don’t need to have knowledge about their location in time and space, while locational predicates must have the temporal and spatial information within them. Swinburne believed that in the case of grue, we should prefer qualitative predicates instead of locational. I think, Swinburne’s objection theory is not complete as well, as for the premise locational predicates might be the correct ones, but for the sake of simplifying it, we prefer to use qualitative ones. (Swinburne, 1968)

The problem, I believe, can be solved if we analyse it within the scopes of natural kinds. Goodman’s approach in defining the word grue is based on contradiction. If a certain object contains two different characteristics that are based on their spatial and temporal information, then the natural kind cannot be considered natural if it is influenced by external forces. It must be free from all kind of forces otherwise it will never reach to its initial natural form. The same logic of the riddle can apply to the statement that “ a triangle is something that has three angles before 2050, and has four angles after 2050”. If triangles are known for having three angle, hence change in their characteristic will eventually change their nature, as well. We won’t accept change in their nature, hence the “new triangle” will become a completely new object for examination. The same can be applied to emeralds and any other object. We acknowledge emeralds for being green (assuming its natural kind) but never blue. If by any chance, emeralds stop being green and change their colour to blue than the object will no longer be an emerald. Chemistry can be a great example, where the natural kind of an element became the basis of the following science. It shows that the natural kind can be exact within the measures of qualitative predicates, simply by the number of protons around their nuclei.

Goodman raised an important issue for illustrating the fallacy of inductive reasoning. However, his created term includes two completely different sides and qualities of the same object. Natural kind cannot be defined two things in a different time, as it must be something that will not change under any circumstances.