The study of how language influences thought and vice versa has a long history in a variety of fields. There are two bodies of thought forming around the debate. One body of thought stems from linguistics and is known as the Sapir–Whorf hypothesis. There is a strong and a weak version of the hypothesis that argue for more or less influence of language on thought. The strong version, linguistic determinism, argues that without language, there is and can be no thought (a largely-discredited idea), and the weak version, linguistic relativity, supports the idea that there are some influences from language on thought. On the opposing side, there are 'language of thought theories', which believe that public language is not essential to private thought though the possibility remains that private thought when infused with inessential language diverges in predilection, emphasis, tone, or subsequent recollection. Those theories address the debate of whether thought is possible without language, which is related to the question of whether language evolved for thought. These ideas are difficult to study because it proves challenging to parse the effects of culture versus thought and of language in all academic fields.

The main use of language is to convey information. It can be used to transfer thoughts from one mind, to another mind, and to modify and explore thoughts within a mind. The bits of linguistic information that enter one person's mind from another cause people to entertain a new thought with profound effects on their world knowledge, inferencing, and subsequent behavior. In the act of speaking, thought comes first, and spoken or written language is an expression that follows. Language has certain limitations, and humans cannot express all that they think. Writing was a powerful new invention because it enabled revision of language and allowed an initial thought to be conveyed, reviewed, and revised before it is expressed.

Language can also be used for thought by framing and modifying thinking with a precision that was not possible without language.

Language of thought theories rely on the belief that mental representation has linguistic structure. Thoughts are "sentences in the head" and so take place within a mental language. Two theories work in support of the language of thought theory. Causal syntactic theory of mental practices hypothesizes that mental processes are causal processes defined over the syntax of mental representations. Representational theory of mind hypothesizes that propositional attitudes are relations between subjects and mental representations. In tandem, these theories explain how the brain can produce rational thought and behavior. All three of those theories were inspired by the development of modern logical inference. They were also inspired by Alan Turing's work on causal processes that require formal procedures within physical machines.

The language of thought hypothesis hinges on the belief that the mind works like a computer, always in computational processes. The theory believes that mental representation has both a combinatorial syntax and compositional semantics. The claim is that mental representations possess combinatorial syntax and compositional semantic—that is, mental representations are sentences in a mental language. Turing's work on physical machines implementation of causal processes that require formal procedures was modeled after these beliefs.

Another prominent linguist, Steven Pinker, developed the idea of a mental language in his book The Language Instinct (1994). Pinker refers to the mental language as "mentalese". In the glossary of his book, Pinker defines mentalese as a hypothetical language that is used specifically for thought. The hypothetical language houses mental representations of concepts such as the meaning of words and sentences.

Different cultures use numbers in different ways. The Munduruku language, for example, has number words only up to five. In addition, it refers to the number 5 as "a hand" and the number 10 as "two hands". Numbers above 10 are usually referred to as "many".

Perhaps the counting system that is the most different from that of modern Western civilisation is the "one-two-many" system used in the Pirahã language in which quantities larger than two are referred to simply as "many". In larger quantities, "one" can also mean a small amount and "many" a larger amount. Research was conducted with the Pirahã by using various matching tasks. The non-linguistic tasks were analyzed to see if their counting system or, more importantly, their language affected their cognitive abilities. The results showed that they perform quite differently from, for example, an English-speaking person who has a language with words for numbers more than two. For example, they represented numbers 1 and 2 accurately using their fingers, but as the quantities grew larger (up to 10), their accuracy diminished. This phenomenon is also called the "analog estimation", as numbers get bigger the estimation grows. Their declined performance is an example of how a language can affect thought and great evidence to support the Sapir–Whorf hypothesis.