NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Outros Autores: | |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Manuscrito (Online) |
Texto Completo: | http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-60452022000200050 |
Resumo: | Abstract Philosophy of language and computer science, despite being very distinct fields, share a great interest in natural language. However, while philosophy has traditionally opted for a formalist approach, computer science has been increasingly favoring probabilistic models. After presenting these two approaches in more detail, we discuss some of their main virtues and limitations. On the one hand, formalist models have trouble in acquiring semantic information from corpora and learning from large amounts of data. Probabilistic approaches, on the other hand, have difficulty in operating with compositionality, in dealing with contrast sets and hierarchical relations, and in distinguishing normative and descriptive views of meaning. We argue that a more fruitful dialogue between philosophers and computer scientists may help to produce a better approach to natural language and stimulate the integration of logical and probabilistic methods. |
id |
UNICAMP-17_6bb5642fd6f306965625bd2780d67498 |
---|---|
oai_identifier_str |
oai:scielo:S0100-60452022000200050 |
network_acronym_str |
UNICAMP-17 |
network_name_str |
Manuscrito (Online) |
repository_id_str |
|
spelling |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCEPhilosophy of LanguageMeaningUseNatural Language ProcessingAbstract Philosophy of language and computer science, despite being very distinct fields, share a great interest in natural language. However, while philosophy has traditionally opted for a formalist approach, computer science has been increasingly favoring probabilistic models. After presenting these two approaches in more detail, we discuss some of their main virtues and limitations. On the one hand, formalist models have trouble in acquiring semantic information from corpora and learning from large amounts of data. Probabilistic approaches, on the other hand, have difficulty in operating with compositionality, in dealing with contrast sets and hierarchical relations, and in distinguishing normative and descriptive views of meaning. We argue that a more fruitful dialogue between philosophers and computer scientists may help to produce a better approach to natural language and stimulate the integration of logical and probabilistic methods.UNICAMP - Universidade Estadual de Campinas, Centro de Lógica, Epistemologia e História da Ciência2022-04-01info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersiontext/htmlhttp://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-60452022000200050Manuscrito v.45 n.2 2022reponame:Manuscrito (Online)instname:Universidade Estadual de Campinas (UNICAMP)instacron:UNICAMP10.1590/0100-6045.2022.v45n2.piinfo:eu-repo/semantics/openAccessPIROZELLI,PAULOCÂMARA,IGOReng2022-05-09T00:00:00Zoai:scielo:S0100-60452022000200050Revistahttp://www.scielo.br/scielo.php?script=sci_serial&pid=0100-6045&lng=pt&nrm=isoPUBhttps://old.scielo.br/oai/scielo-oai.phpmwrigley@cle.unicamp.br|| dascal@spinoza.tau.ac.il||publicacoes@cle.unicamp.br2317-630X0100-6045opendoar:2022-05-09T00:00Manuscrito (Online) - Universidade Estadual de Campinas (UNICAMP)false |
dc.title.none.fl_str_mv |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
title |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
spellingShingle |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE PIROZELLI,PAULO Philosophy of Language Meaning Use Natural Language Processing |
title_short |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
title_full |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
title_fullStr |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
title_full_unstemmed |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
title_sort |
NATURAL LANGUAGE AT A CROSSROADS: FORMAL AND PROBABILISTIC APPROACHES IN PHILOSOPHY AND COMPUTER SCIENCE |
author |
PIROZELLI,PAULO |
author_facet |
PIROZELLI,PAULO CÂMARA,IGOR |
author_role |
author |
author2 |
CÂMARA,IGOR |
author2_role |
author |
dc.contributor.author.fl_str_mv |
PIROZELLI,PAULO CÂMARA,IGOR |
dc.subject.por.fl_str_mv |
Philosophy of Language Meaning Use Natural Language Processing |
topic |
Philosophy of Language Meaning Use Natural Language Processing |
description |
Abstract Philosophy of language and computer science, despite being very distinct fields, share a great interest in natural language. However, while philosophy has traditionally opted for a formalist approach, computer science has been increasingly favoring probabilistic models. After presenting these two approaches in more detail, we discuss some of their main virtues and limitations. On the one hand, formalist models have trouble in acquiring semantic information from corpora and learning from large amounts of data. Probabilistic approaches, on the other hand, have difficulty in operating with compositionality, in dealing with contrast sets and hierarchical relations, and in distinguishing normative and descriptive views of meaning. We argue that a more fruitful dialogue between philosophers and computer scientists may help to produce a better approach to natural language and stimulate the integration of logical and probabilistic methods. |
publishDate |
2022 |
dc.date.none.fl_str_mv |
2022-04-01 |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-60452022000200050 |
url |
http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-60452022000200050 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
10.1590/0100-6045.2022.v45n2.pi |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
text/html |
dc.publisher.none.fl_str_mv |
UNICAMP - Universidade Estadual de Campinas, Centro de Lógica, Epistemologia e História da Ciência |
publisher.none.fl_str_mv |
UNICAMP - Universidade Estadual de Campinas, Centro de Lógica, Epistemologia e História da Ciência |
dc.source.none.fl_str_mv |
Manuscrito v.45 n.2 2022 reponame:Manuscrito (Online) instname:Universidade Estadual de Campinas (UNICAMP) instacron:UNICAMP |
instname_str |
Universidade Estadual de Campinas (UNICAMP) |
instacron_str |
UNICAMP |
institution |
UNICAMP |
reponame_str |
Manuscrito (Online) |
collection |
Manuscrito (Online) |
repository.name.fl_str_mv |
Manuscrito (Online) - Universidade Estadual de Campinas (UNICAMP) |
repository.mail.fl_str_mv |
mwrigley@cle.unicamp.br|| dascal@spinoza.tau.ac.il||publicacoes@cle.unicamp.br |
_version_ |
1748950066252480512 |