with illustrations from Conan Doyle’s stories about Sherlock Holmes

Michael Stubbs

University of Trier

The Prague Linguistic Circle, which met at Charles University in the 1920s and 1930s, included Czech linguists (Vilém Mathesius and Josef Vachek), Czech and Austrian-Czech literary scholars (Jan Mukařovský and René Wellek), and Russian émigré linguists (Roman Jakobson and Nikolai Trubetzkoy). In a cultural and political history of the group, Toman (1995) discusses their aims of “redefining the nature of linguistics and literary studies”.

The group tackled problems which have still not been solved, including the nature of literary language and the relation of author to text. Two significant quotes are:

“The function of poetic language is the maximum foregrounding of the utterance” (Mukařovský 1932).

“We are not interested in the individual psychology of the novelist, but in his novel as an objective social fact” (Jakobson 1943).

The concept of “intertextuality” suggests an empirical approach to these problems. It is used of cases where units of meaning in one text refer to units of meaning in another text, and implies that the meaning of a text does not depend on (the intention of) the author, but on how readers interpret relations between texts.

The term “intertextuality” was introduced in the 1960s, and in the last fifty years a huge literature has appeared: JSTOR returns over one thousand items with “intertext*” in the title (almost all by literary and cultural theorists, few by linguists). However, this intensive discussion has arguably made only minimal conceptual progress. Since the concept has failed to develop significantly over such a long period, perhaps it should be abandoned in favour of something more worthwhile.

Alternatively, we can argue as follows. First, intertextuality can now be studied by new kinds of corpus data and methods, which were previously not available to literary and cultural scholars. Second, the concept of intertextuality is logically related to other concepts, such as reference, semantic units, paraphrase, and evaluative language. Third, a concept is seen to be significant when it is connected, in a natural way, to a complex of other ideas: if such connections can be made, this often leads to intellectual progress.

The essential empirical question is: Can corpus methods reliably identify intertextual references? The essential conceptual question is: What is the logical relation between intertextuality and meaning?

Conan Doyle’s stories about Sherlock Holmes (published 1887 to 1927) provide an ideal corpus for studying intertextuality. They refer to identifiable texts (e.g. earlier detective stories) and allude to contemporary ideas, both scientific (e.g. the value of observational data) and pseudo-scientific (e.g. so-called “criminal anthropology”).

The stories were clearly influenced by other fictional and non-fictional texts which characterize the intellectual and social world of the late 1800s. These texts are “objective social facts”, but relations between texts change over time, and so therefore does the meaning of the texts. Corpus data on these relations could help to solve problems of language, literature and culture which were posed by the Prague Linguistic Circle over eighty years ago. Corpus et orbis …


Jakobson, Roman. 1943. Unpublished lecture. Quoted by Toman 1995: 247-48.

Mukařovský, Jan. 1932. Standard language and poetic language. In P. Garvin ed. 1964. A Prague School Reader on Esthetics, Literary Structure and Style. Georgetown UP.

Toman, Jindřich. 1995. The Magic of a Common Language: Jakobson, Mathesius, Trubetzkoy, and the Prague Linguistic Circle. MIT Press.