I just read an extract of David Lewis's Counterfactuals and he claims there in a footnote on page 90 that there are at least beth_2 possible worlds. He also claims in the very same footnote that "[i]t can easily be shown that this is the number of ersatz worlds in Quine’s original construction." (Lewis 1973, p. 90) Would anyone know if there is literature on this statement (maybe even a proof of this proposition) and more generally on the cardinality of the logical space? Thank you in advance for your help!

  • 1
    Since there're ℶ0 number of sentences of a literal language, there're at most ℶ1 number of sets of such sentences. Clearly it's not enough to correspond to all conceivable ersatz handles of PWs since for example the value of some physical quantities can be ranged over a continuum as a real number or possibly there're new sentences obtained in some PW where the existing language simply cannot express, thus it must be at least the next large cardinal... Feb 18, 2022 at 2:11

1 Answer 1


It seems strange to me. Firstly, we would need to specify what type of language is D. Lewis considering. In addition, it is necessary to specify what is a word within such language.

To simplify, a language is based on combinations of letters of a certain alphabet (written languages have finite alphabets and spoken languages have a finite number of distinct phonemes). With a finite alphabet, a word is a finite sequence of symbols of the alphabet, the possible descriptions in such a language are infinite. Specifically, the inifnite set of finite sequences of words in such a language has cardinality ℵ0 = ℶ0. The the beth numbers are defined recursively as:

n+1 = 2n

So, this number is equivalent to the possible infinite collections of sentences in the specified language (but these sets are, themselves not expressible by means of the language). This situation appears in formal languages. For example, using the current mathematical notation we can specify an infinite collection of entities, with cardinality ℶ0 (equals ℵ0). This means for example, that the vast majority of real numbers cannot be "referred to" using the conventional notation, because there are ℶ1 real numbers, but only ℶ0 possible expressions to denote the entities and ℶ0 < ℶ1.

On the other hand, ℶ2 = 21 is another amazingly large infinite number/cardinal, which of course, includes many more entities than can be named with ordinary formal language.

  • Worlds, not words.
    – benrg
    Feb 18, 2022 at 1:00
  • 1
    As @benrg suggests, it looks like you misread the question. The question is about possible worlds, not possible words. Suggest you delete this answer as not responsive to the question. Feb 27, 2022 at 4:17
  • I think Davius might be working off the "worlds as maximal sentence sets" way of talking about possible worlds. Or if not, this answer could be edited accordingly (and quickly/easily) to try to correlate "the number of words" with "the number of sentences that can be made from those words" and go on to "the number of worlds that can be traced back to those recombinations of words." May 3, 2022 at 18:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .