The SLX Corpus of Classic Sociolinguistic Interviews comprises 8 sociolinguistic interviews with a total of 9 speakers, conducted in the 1960s and 70s. All of the interviews are conducted by William Labov or by one of his students. Labov notes that these interviews are not classic in the sense that they form part of a systematic sociolinguistic study of the speech community. Only the Rose B. interview is, drawn from the NYC study. And from the point of view of methodology -- what makes the other interviews good -- it is defective. It still bears the hallmarks of the dialectological format from which it was drawn. The others are all exploratory interviews without any basis in systematic sampling or design.

What makes these interviews classic is that they represent classic solutions to the problems of achieving cross-cultural contact, reducing the effect of the Observer's Paradox and approximating the vernacular of every-day life. Most importantly, they are interviews with extraordinarily gifted, memorable and fluent speakers. They contain a very rich store of narratives in response to the classic set of sociolinguistic interview questions. Several proved useful for the acoustic analysis of Labov, Yaeger & Steiner 1972 (Henry G., Kathy D., Jerry T.); and others provided materials for Labov's work on narrative analysis (Henry G., Bobbie A., Adolphus G., Joe D.).

These particular interviews have also been targeted for inclusion in this corpus because of the sound quality. All interviews were recorded on a Nagra III or IVS with Sennheiser dynamic microphones. Finally, these particular speakers have been selected because publication of the audio data and corresponding transcripts and annotations does not violate any agreement the interviewer made with the speakers regarding data distribution.

The data reflects a broad spectrum of speaking styles, including spontaneous speech, narratives, responses and formal linguistic tasks. The interviews touch on a multitude of topics, and corpus users should note that the language of the interviews represents the uncensored opinions of the speakers, reflecting their every day concerns and personal histories.

Taken as a whole, the speakers exemplify a wide variety of regional and social dialects. Demographic information for each main speaker in the corpus is displayed in the table below.

To create the current corpus, interview sessions were digitized from the original open reel tapes onto DAT/disk at 16bit, 44KHz sampling. The monaural signal was passed through 2 channels at levels differing by 20% to capture the best digital copy in a single pass. Technicians monitored the recording process, and adjusted for sustained changes in speech levels. The digital files show no significant clipping in the digital domain.

The final SLX corpus publication includes 17 speech files downsampled to 22050Hz, in single-channel MS WAV (RIFF) format, for a total of 575 minutes (~ 1.5GB) of speech.

In the next stage, the audio files were processed to produce a single transcript file for each speaker. The target speaker was distinguished from the interviewer, other speakers, silence and noise. A time-aligned transcript was then created for each speaker. Audio segmentation proceeded in two passes. The first identified basic utterance boundaries. Sentence or phrase boundaries, significant pauses and breath groups were identified and marked in the second pass.

Once individual utterance-length segments had been created for each speaker, transcription began. The transcripts are written in standard orthography with normal punctuation and standard spelling. No attempt has been made to correct speakers' pronunciation or grammar, and problematic or incomprehensible sections are transcribed using special conventions. Unintelligible speech or doubtful best guesses, for example, are enclosed in ((double parentheses)). Non-standard lexical items are marked with special symbols. All transcripts have undergone at least two further passes: the first to catch errors and revisit unintelligible sections, and the second to allow a native speaker of British English to check for errors based on unfamiliarity with cultural items, slang or pronunciation.

This publication includes a number of annotation tools that were developed to support browsing and exploration of the SLX Corpus, as well as those used in the creation of the corpus. The current release includes a pre-compiled package of these tools for Windows users. We plan to create a pre-compiled package for i386 Linux users in the future.

The SLX corpus is an example of a multi-community sociolinguistic corpus. Containing classic interview material in the Labovian tradition, it is a valuable teaching tool for linguists. The recordings demonstrate successful interviewing techniques, the sound quality is high, and the digitization, segmentation and transcription of the data represent best practice in these areas. The variable survey highlights over 150 sociolinguistic variables attested in the corpus and suggests avenues for further research. Most importantly, the SLX corpus represents an important step in creating infrastructure, tools and data to support a collaborative digital methodology for the quantitative study of linguistic variation in society.

The annotation tools included within the current publication have been optimized for browsing the SLX Corpus. Future tool releases will integrate the segmentation and transcription features of MultiTrans and the annotation features of DASLTrans into a single tool, allowing users to go from segmentation to transcription to coding of variables within a single system.