Verb frames in Early Irish law texts

These files in excel and csv format contain a dataset of the valency frames of verbs attested in six Old Irish law texts (1st release, 692 frames).

The texts are as follows:

BB        Bechbretha. Al Old Irish law-tract on bee-keeping, ed. by Thomas Charles-Edwards & Fergus Kelly. Dublin, DIAS 1983.

BC        Bretha Comaithcheso. An Old-Irish Law Tract on Neighbouring Farms, ed. by Thomas Charles-Edwards. Dublin, DIAS 2022.

BNT     Bretha Nemed Toísech, ed. by Liam Breatnach, "The First Third of Bretha Nemed Toísech", Ériu 40 (1989): 1-40.

CG       Críth Gablach, ed. by Daniel Binchy. Dublin, DIAS 1941.

MD       Marriage Disputes: a fragmentary Old Irish Law-Text, ed. by Fergus Kelly. Dublin, DIAS 2014 (only Old Irish text).

UnR     Uraicecht na Ríar. The poetic grades in Early Irish Law, ed. by Liam Breatnach. Dublin, DIAS 1987.

The frames are given in the format established by the Valency Patterns Leipzig database (ValPaL https://valpal.info/) and Pavia Verbs Database (PaVeDa https://paveda.unipv.it/), although they have been slightly simplified for the purposes of this database, according to the following principles:

·       each argument is numbered and given both a semantic microrole (e.g. the first argument of the active forms of téit ‘go’ is a ‘goer’) and a morphosyntactic coding set (case or preposition+case, e.g. 1-nom corresponds to the ‘goer’ of téit); when word order is the only relevant coding, the symbol > has been used (this occurs only in the costruction where a verbal noun in the dative case is preceded by one of its arguments and do)

·       V.PASS is used for frames with passive verb forms. The frames with the passive that contain a nominative argument include all instances where 1st and 2nd persons are coded as infixed (object) pronouns.

·       V’ is used for verbal noun forms. The frames for verbal nouns are included as long as they have a corresponding finite form (to whose lemma they are connected) and as long as they appear with at least one overt argument/adjunct (for example nouns with verbal noun behaviour such as cáe, coí ‘coshering visit’ or slógad ‘hosting’ that occur in CG are excluded since they do not have a corresponding finite verb; suide in the instance gaimniu suidi ‘couch for sitting’ at CG 404 is also excluded, because there is no overt argument of suide).

·       agreement of the verb with nominative arguments is not specified as in the ValPaL/PaVeDa frames (e.g. V.subj[1] 1-nom), but taken for granted. In case of frames where there is no nominative argument (impersonal constructions with active forms), the verb form is given as V.3SG, e.g. V.3SG di+1-dat do+2-dat is the frame for níbi di chaind airchetal dó ‘he is not from a family which practices poetry’, UnR 10.

All verbs occurring in the six texts are included but not all copular constructions (verb lemma is) are included. Only some specific copular constructions, which are not simply formed by a nominal or adjectival predicate and a nominal subject, and whose meaning is not plainly compositional, have been included, somewhat arbitrarily.

The first sheet (meanings and microroles) contains the frames (each row represents a frame) and the following information (in columns): meaning of the verb with each frame, verb lemma, valency frame, microrole (mr) and coding set (cs) for each numbered argument in the frame. The microroles in ValPaL and PaVeDa are pre-assigned roles but in this database they have been created ex novo following similar conventions to name them. Phrases that do not appear to be arguments but adjuncts, as they are not required by the verb, or do not frequently occur with it, have often been included in order to collect data that may give clearer proofs for or against their argumental status, finding out whether they occur with the same semantic role in different contexts and with different verbs. Optional arguments of a single frame, i.e. a frame where the other arguments may occur without that argument but with the same meaning, are put within parentheses.

The second sheet (examples) contains at least one example for each frame, and includes the following information (in columns): verb lemma and valency frame as in the first sheet, (primary) edited text of the example, English translation, optional comments, source and locus. The translations given are usually taken from edited sources but I occasionally suggest alternative interpretations. The comment column also contains information about occurrence in “poetry” (which includes different kinds of rhyming verse and alliterative roscada). In a forthcoming release I plan to add analysed text and morphological glosses for a convenient number of examples. When many instances of a frame occur in a text or in different texts, not all instances are given in full, but all loci are listed in the source and loci column.

In some cases the analysis of the frame could not be established with certainty: the corresponding row (or cell, when only details such as case are doubtful) is coloured in yellow.

The third sheet contains the abbreviations used in the frames as well as in the comments and for sources.

Each information is of course searchable, so that different frames for a verb lemma or verb meaning, or different coding sets for similar microroles might be searched for.

For the time being (1st release, January 2026), no attempt has been made to identify basic coding frames, i.e. those valency patterns that appear to be more frequent and from which other frames might be considered derived (i.e. alternations in ValPaL and PaVeDa terms). In a sense, basic coding frames are the constructions that one would expect to find at the beginning of a dictionary entry for a verb. This choice is left for future plans, to a stage when data collection has reached a more advanced status and data may be considered large enough.

The design of this database has been presented by its author at the Fourth European Symposium in Celtic Studies, Bonn, on August 28th 2025 (https://www.iaak.uni-bonn.de/keltologie/en/aktuelles/tagungen).

All comments are welcome: please write to [e-mail]