One of the best earlier answers to this question came from Dell Hymes, who in his introduction to Toward ethnographies of communication said that formalization should be a “preparation of better means to anthro-pological ends” (Hymes 1964:1). This idea was earlier expressed by others, like Sapir, who said that linguists “must become increasingly concerned with the many anthropological, sociological, and psychological problems which invade the field of language” (Sapir 1929:166).
These statements make clear that language has more to do with communication and society than it has to do with an inherent property of humans. Because in the case that we adopted the second posture and thought of language as a system that is wired into humans’ brains or an inborn quality that distinguishes us from other animals, many questions related to its evolutionary origin would arise. Language could be said to have a system that is ingrained into the human brain, but not from birth but through social practice. It is important also to remark that language “has” a system that allows for better communicative efficiency, but that “it is not” a system. Proof of it are the numerous pidgins and sign languages whose creations and developments have been recorded. Those languages never started as a system, but as a chaotic instinct for communication. The system was added by the users to gain efficiency. But a clearer and much commoner example that language is not a system is the implementation of grammar upon a language. We can see that people can acquire and use a language without any knowledge of grammar. Ungrammatical language users may eventually correct themselves to gain efficiency while conveying their messages, but there are also other users: poets and writers in general, who may opt to be less grammatical just to add complexity to their thoughts. But the mere fact that grammar is a subject at schools is proof enough that it is not inherent to language. Grammar, or a system, are requirements to identify a language as Spanish, English or Polish, but they are not primordial for communication.
So if we are looking for an inborn quality in human beings, we will be disappointed to learn that language is not a quality that distinguishes us from animals, at least not genetically. The need for communication, which is the basis for language, as well as for other kinds of communication seen in animals in general, is the only wired ability we have. The quality that may distinguish us from animals is our “culture,” which is greater than theirs. Animals may have cultures too, which they pass down from generation to generation. It has been proven that animals can teach and learn. But humans’ culture is greater, and only this is what accounts for language. The arbitrariness of languages is proof of it. What may at the beginning have seemed like instinctive reactions to the environments: onomatopoeic sounds and imitative gestures, have become totally arbitrary sounds in every proper language. Then, the amount of concepts and words is what accounts for a system, without which those languages would not have survived.
Hymes has given a good example of the communicative function of language: its semantic use. He says that “Although anthropologists have sometimes talked of the use of language ‘merely’ as a tool of communication, and of the categorizing of experience as if it were a superior category, the role of a language as a device for categorizing experience and its role as an instrument of communication cannot be so separated, and indeed, the latter includes the former” (Hymes 1964:20). The use of language to describe things and events therefore belongs to the communicative function of the language, according to Hymes. That is why he affirms that a linguistic analysis “must take as context a community, investigating its communicative habits as a whole, so that any given use of channel and code takes its place as but part of the resources upon which the members of the community draw” (Hymes 1964:3). Human language then is just a combination of the instinct for communication present in all animals and human culture. Therefore language as such must be also attributed to every creature capable of creating culture. If we look at dogs and other pets, we can see that they have a very systematic way of communicating their emotions and needs. This must be called language, provided it is not random. Then a language could be defined as: A systematic means of communication, but still it would not be a “system,” but a “means of communication.” That accounts for languages that are more systematic than others, languages whose grammar is more strict and languages that are more semantically flexible than others.
The last thing to remark while analyzing the origin of language is the artificiality of writing. Writing is the best example of linguistic culture, because it is totally artificial. This is not so evident in phonetic languages that use phonetic symbols to create words, but it is more visible in logographic or logosyllabic languages, like for instance Chinese. Here writing is totally dissociated from speaking and is totally independent from it; therefore they can evolve separately due to different levels or forms of cultivation. But the relationship between speech and writing shows how the relationship between communication and language works. Both are independent processes that may happen simultaneously or unilaterally and that share a common element. And although both writing and language may be thought to be determined by speech and communication respectively, they are independent because they are basically a cultural product, while speech and communication are relatively more instinctive.
1964 Toward Ethnographies of Communication, in American Anthropologist, New Series, Vol. 66, No. 6, pp. 1-34. Blackwell Publishing.
1949 The status of linguistics as a science. Language 5: 207-214. Reprinted in Selected writings of Edward Sapir, David G. Mandelbaum, ed. Berkeleya nd Los Angeles, Universtiy of California P ress, 1949.