The Brooklyn Corpus: Syntactic Annotations, File Formats, and Search Tools

The syntactic annotations of the Brooklyn Corpus enable the users to pose and answer questions about word order, constituent order, abstract structure, and syntactic and morphological characteristics of the texts in the corpus. The annotations are general-purpose and as theory-neutral as possible, while still incorporating the insights of modern linguistic theory, and they can be used by scholars with widely varying research interests. But it must be emphasized that the annotations should in no way be regarded as the implementation of a structural analysis, and that the annotation schemes were developed primarily as a tool for the investigation of the structure of earlier stages of English.

The syntactic annotations mark constituents, both clausal and non-clausal, by labelled brackets, with some relations marked by empty categories. The structure assigned to a sentence by the labelled bracketing can be quite complex, but it is not a complete syntactic analysis: the function of the bracketing is not to assign a structure to Old English sentences but rather to facilitate searches.

Below is an example of an annotated sentence from the Brooklyn Corpus, with its gloss:

( [rt +ta/RT ] [vt com/VT ] [nn sum/JJN wydewe/NNN , 1[L [wh-1 seo/PDN ] %co% %nn-1% [vt w+as/ET ] [vn geciged/VN ] 2[S %nn-1% [nn Euthicia/NMN ] S]2 L]1 , ] [pp betwux/PP o+drum/JJD mannum/NND ] [pp to/PP +t+are/PDD m+aran/JJD byrigene/NND ] ... ) (AELIVE,I,210.5)

then came some widow, that was called Eutychia, between other persons to that famous grave ...

The sentence begins with the temporal adverb '+ta', followed by the finite verb 'com', followed by the subject 'sum wydewe', which is modified by the relative clause 'seo w+as geciged Euthicia'; followed by two preposition phrases 'betwux o+drum mannum' and 'to +t+are m+aran byrigene'. The annotation is described in detail in the manual accompanying the corpus.

• Format 1, with extension .o2/.o23/.o24/.o3/.o34: the largest version of the corpus, with a header block and three sections for each token: 1) the Old English data with syntactic and part-of-speech annotation, 2) the gloss, and 3) the morphological annotation. This format can be searched only by search functions within word-processing programs and editors.

• Format 2, with extension .oe: the Old English data with syntactic and part-of-speech annotation only. This format can be searched by word-processing programs and by the PERL scripts supplied with the Brooklyn Corpus.

• Format 3, with extension .unpacked: the Old English data unpacked, with one clause per token. This format can be searched by word-processing programs and by the PERL scripts supplied with the Brooklyn Corpus.

• Format 4, with extension .psd: the Old English data in a format suitable only for searching with CorpusSearch, a powerful search engine developed by Beth Randall for the second edition of the Penn-Helsinki Parsed Corpus of Middle English (PPCME2). For information about PPCME2 and CorpusSearch, and for order forms, click here.