The York—Toronto—Helsinki parsed corpus of old english prose

Taylor, Ann; Warner, Anthony; Pintzuk, Susan; Beths, Frank

doi:10.1057/9780230223202_9

Cited by 138 publications

(149 citation statements)

References 7 publications

Supporting

Mentioning

139

Contrasting

Unclassified

Order By: Relevance

“…Surprisingly, however, we show that the two filtering constraints on the input are crucially involved in explaining the change in Old English from a strongly object-verb (OV) distribution to a strongly verb-object (VO) distribution between 1000 A.D. and 1200 A.D (YCOE corpus : Taylor, Warner, Pintzuk, & Beths, 2003;PPCME2 corpus: Kroch & Taylor, 2000). These filters on the input during learning must be in place for a model to simulate the correct rate of change that matches the Old English population's rate.…”

mentioning

confidence: 83%

“…Between 1000 A.D. and 1150 A.D., the distribution in the Old English population mostly consisted of OV order utterances (7a), whereas the distribution in the population at 1200 A.D. mostly consisted of VO order utterances (7b; YCOE corpus: Taylor et al, 2003;PPCME2 corpus: Kroch & Taylor, 2000 Unambiguous data for OV word order correlate with observable data of the following types in Old English: The tensed Verb appears at the end of the clause (8a) or the nontensed verb remains in the postobject position, whereas the tensed auxiliary moves (8b). Given V2 (where the tensed verb moves to the second phrasal position in the sentence and some other phrase moves to the first position), a simple subject tensedverb object utterance could be parsed with either OV or VO order.…”

Section: Old English Changementioning

confidence: 99%

See 1 more Smart Citation

Input Filtering in Syntactic Acquisition: Answers From Language Change Modeling

Pearl

Weinberg

2007

Language Learning and Development

View full text Add to dashboard Cite

We use historical change to explore whether children filter their input for language learning. Although others (e.g., Rohde & Plaut, 1999) have proposed filtering based on string length, we explore two types of filters that assume richer linguistic structure. One presupposes that linguistic utterances are structurally highly ambiguous and focuses learning on unambiguous data (Dresher, 1999;Fodor, 1998b;Lightfoot, 1999). The second claims that children learn only from matrix clauses (Lightfoot, 1991), defining simplicity in a structural manner. We assume that certain language changes occur via mismatches during acquisition. This allows us to use patterns of change to demonstrate that filtering restrictions are necessary to model language learning. Viewing language change as a result of mismatches during learning thus constrains the learning algorithm itself.Scientists have attributed asymmetries between typical adult and child knowledge of language to filtering of the input data. For instance, Morgan (1986) claimed that children, perhaps because of general cognitive restrictions on the complexity of data that they can handle, might restrict attention to simple subparts of utterances. Elman (1993) and others have suggested that this constraint might be motivated by architectural considerations of the underlying language analyzer, modeled as a simple recurrent network-but see Rohde and Plaut (1999) for simulations that cast this assertion into doubt.A ubiquitous property of natural language is its rampant structural ambiguity. Particular strings can be analyzed in multiple ways given all of the grammatical LANGUAGE LEARNING AND DEVELOPMENT, 3(1), 43-72 Copyright © 2007, Lawrence Erlbaum Associates, Inc. Correspondence should be addressed to Lisa Pearl, Linguistics Department, 1401 Marie Mount Hall, University of Maryland, College Park, MD 20742. E-mail: llsp@umd.edu, weinberg@ umiacs.umd.edu options available cross-linguistically. The wrong analysis of such a string can lead the child to include a rule in the underlying grammar that is wrong for the language as a whole. Subsequent revision of the underlying grammar to exclude this rule can then be costly, so this is a serious concern.Dresher (1999), Lightfoot (1991), and Fodor (1998b have proposed learning strategies that bias children away from these potentially misleading cases, using a structurally based definition of simplicity to filter the data. Unfortunately, it is nearly impossible to test any filtering proposal in a natural setting. For logistical and ethical reasons, one cannot simply expose a child to an unnaturally restricted dataset during the critical period of language learning to observe the effect of that restriction on the normal course of language acquisition.Modeling language change, however, can offer a graceful solution to this predicament if we assume that certain types of change result from a misalignment of the child's hypothesis and an adult's analysis of the same data (Lightfoot, 1991(Lightfoot, , 1999. Language change models incorpor...

show abstract

mentioning

confidence: 83%

Section: Old English Changementioning

confidence: 99%

Input Filtering in Syntactic Acquisition: Answers From Language Change Modeling

Pearl

Weinberg

2007

Language Learning and Development

View full text Add to dashboard Cite

show abstract

“…ME and OE The Penn-Helsinki Parsed Corpus of Middle English (PPCME2) 1 and the York-TorontoHelsinki Parsed Corpus of Old English Prose (Taylor et al, 2003b, YCOE) use a variant of the PTB annotation schema (Taylor et al, 2003a). YCOE contains the full West Saxon Gospel, but PPCME2 contains only a small fragment of a Wycliffite gospel of John, the ME data is thus complemented with parts of Genesis (G) and Numbers (N).…”

Section: Languages and Corpus Datamentioning

confidence: 99%

Diachronic proximity vs. data sparsity in cross-lingual parser projection. A case study on Germanic

Sukhareva¹,

Chiarcos²

2014

Proceedings of the First Workshop on Applying NLP Tools to Similar Languages, Varieties and Dialects

View full text Add to dashboard Cite

For the study of historical language varieties, the sparsity of training data imposes immense problems on syntactic annotation and the development of NLP tools that automatize the process. In this paper, we explore strategies to compensate the lack of training data by including data from related varieties in a series of annotation projection experiments from English to four old Germanic languages: On dependency syntax projected from English to one or multiple language(s), we train a fragment-aware parser trained and apply it to the target language. For parser training, we consider small datasets from the target language as a baseline, and compare it with models trained on larger datasets from multiple varieties with different degrees of relatedness, thereby balancing sparsity and diachronic proximity.Our experiments show (a) that including related language data to training data in the target language can improve parsing performance, (b) that a parser trained on data from two related languages (and none from the target language)can reach a performance that is statistically not significantly worse than that of a parser trained on the projections to the target language, and (c) that both conclusions holds only among the three most closely related languages under consideration, but not necessarily the fourth.The experiments motivate the compilation of a larger parallel corpus of historical Germanic varieties as a basis for subsequent studies.

show abstract

“…Modern 19th CS 21K (Kroch et al, 2010) British 18th CS 32K (Kroch et al, 2010) Early 17th CS 22K (Kroch et al, 2004) Modern 16th CS 21K (Kroch et al, 2004) Middle 14th CS 66K (Kroch and Taylor, 2000) Old 10th CS 78K (Taylor et al, 2003b) DS 7K (Haug and Jøhndal, 2008) Icelandic Middle 16th CS 40K (Rögnvaldsson et al, 2012) High Early Mod.16th CS 27K (Light, 2013) Table 2: Verse-aligned texts in the Germanic parallel Bible corpus (parentheses indicate marginal fragments with less than 50,000 tokens) tiple versions of the same passage in the same language can be provided.…”

Section: Englishmentioning

confidence: 99%

New Technologies for Old Germanic. Resources and Research on Parallel Bibles in Older Continental Western Germanic

Chiarcos

Sukhareva

Mittmann

et al. 2014

Proceedings of the 8th Workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities (LaTeCH)

View full text Add to dashboard Cite

We provide an overview of on-going efforts to facilitate the study of older Germanic languages currently pursued at the Goethe-University Frankfurt, Germany.We describe created resources, such as a parallel corpus of Germanic Bibles and a morphosyntactically annotated corpus of Old High German (OHG) and Old Saxon, a lexicon of OHG in XML and a multilingual etymological database. We discuss NLP algorithms operating on this data, and their relevance for research in the Humanities.RDF and Linked Data represent new and promising aspects in our research, currently applied to establish cross-references between etymological dictionaries, infer new information from their symmetric closure and to formalize linguistic annotations in a corpus and grammatical categories in a lexicon in an interoperable way.

show abstract

The York—Toronto—Helsinki parsed corpus of old english prose

Cited by 138 publications

References 7 publications

Input Filtering in Syntactic Acquisition: Answers From Language Change Modeling

Input Filtering in Syntactic Acquisition: Answers From Language Change Modeling

Diachronic proximity vs. data sparsity in cross-lingual parser projection. A case study on Germanic

New Technologies for Old Germanic. Resources and Research on Parallel Bibles in Older Continental Western Germanic

Contact Info

Product

Resources

About