{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_52975"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_52975","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"From Language Models to Moral Machines : Mechanism and Generalization","abstract":"Thesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2026","abstract_html":"Thesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2026","abstract_has_math":false,"creators":["Liu, Guangliang"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Johnson, Kristen","Wang, Rongrong","Tan, Pan-Ning","Kordjamshidi, Parisa"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T18:59:34Z","subjects":["Computer science"],"languages":["eng"],"rights":["Attribution-ShareAlike 4.0 International"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:52975","local:Liu_grad.msu_0128D_21948"],"render_values":[{"text":"etd:52975","href":null,"code":true},{"text":"local:Liu_grad.msu_0128D_21948","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/h89k-9r20","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/h89k-9r20","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_52975","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Johnson, Kristen","Wang, Rongrong","Tan, Pan-Ning","Kordjamshidi, Parisa"]},{"key":"dc:creator","label":"Author","values":["Liu, Guangliang"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["text","Electronic dissertations"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Computer science"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution-ShareAlike 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:52975","local:Liu_grad.msu_0128D_21948","https://doi.org/doi:10.25335/h89k-9r20"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2026","Humans make daily decisions based on moral considerations, ranging from stopping at a traffic sign to reporting experimental results in publications.Machine morality seeks to enable machines to exhibit human-like morality, and it has been a long-standing research challenge in the fields of artificial intelligence, linguistics, natural language process, etc. Humans acquire morality from social experience and machines, e.g., Large Language Models (LLMs), acquire knowledge from texts. The fundamental differences between humans and machines make it an open challenge to advance machine morality. LLMs excel at modeling distributional semantics which argues that two linguistic units share similar meanings if they occur in similar contexts. For machines to exhibit human-like morality, they must go beyond distributional semantics and develop pragmatic competence. For instance, distributional semantics fails to capture cases like ``Never call your uncle by his name,'' which may be acceptable in some East Asian cultures but considered inappropriate in others, as its interpretation depends on social context—that is, on the pragmatics of morality. In light of this gap between distributional semantics and pragmatics, my research endeavors to uncover the mechanisms through which LLMs acquire and utilize morality, and to develop approaches for advancing machine morality. I draw on techniques and theories from statistical learning theory, linguistics, and social sciences, focusing on three research questions: (1) What are the limitations of modeling morality solely through distributional semantics? (2) How do LLMs process the pragmatics of morality? (3) How to bridge the gap between distributional semantics and the pragmatics of morality? By addressing these research questions, this dissertation finds that linguistic cues relevant to morality remain beyond the distributional semantics captured by LLMs, which therefore rely on distributional patterns to approximate pragmatics. To address this limitation, this dissertation proposes a pragmatic inference method that bridges the gap between distributional semantics and the pragmatics of morality, significantly improving generalization in moral reasoning tasks while demonstrating flexibility and robustness across diverse moral discourses.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","1 online resource (147 p.) : ill."]},{"key":"dc:title","label":"Title","values":["From Language Models to Moral Machines : Mechanism and Generalization"]}]}],"canonical_facts":{"dc:contributor":["Johnson, Kristen","Wang, Rongrong","Tan, Pan-Ning","Kordjamshidi, Parisa"],"dc:creator":["Liu, Guangliang"],"dc:date":["2026"],"dc:description":["Thesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2026","Humans make daily decisions based on moral considerations, ranging from stopping at a traffic sign to reporting experimental results in publications.Machine morality seeks to enable machines to exhibit human-like morality, and it has been a long-standing research challenge in the fields of artificial intelligence, linguistics, natural language process, etc. Humans acquire morality from social experience and machines, e.g., Large Language Models (LLMs), acquire knowledge from texts. The fundamental differences between humans and machines make it an open challenge to advance machine morality. LLMs excel at modeling distributional semantics which argues that two linguistic units share similar meanings if they occur in similar contexts. For machines to exhibit human-like morality, they must go beyond distributional semantics and develop pragmatic competence. For instance, distributional semantics fails to capture cases like ``Never call your uncle by his name,'' which may be acceptable in some East Asian cultures but considered inappropriate in others, as its interpretation depends on social context—that is, on the pragmatics of morality. In light of this gap between distributional semantics and pragmatics, my research endeavors to uncover the mechanisms through which LLMs acquire and utilize morality, and to develop approaches for advancing machine morality. I draw on techniques and theories from statistical learning theory, linguistics, and social sciences, focusing on three research questions: (1) What are the limitations of modeling morality solely through distributional semantics? (2) How do LLMs process the pragmatics of morality? (3) How to bridge the gap between distributional semantics and the pragmatics of morality? By addressing these research questions, this dissertation finds that linguistic cues relevant to morality remain beyond the distributional semantics captured by LLMs, which therefore rely on distributional patterns to approximate pragmatics. To address this limitation, this dissertation proposes a pragmatic inference method that bridges the gap between distributional semantics and the pragmatics of morality, significantly improving generalization in moral reasoning tasks while demonstrating flexibility and robustness across diverse moral discourses.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."],"dc:format":["application/pdf","1 online resource (147 p.) : ill."],"dc:identifier":["etd:52975","local:Liu_grad.msu_0128D_21948","https://doi.org/doi:10.25335/h89k-9r20"],"dc:language":["eng"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["Attribution-ShareAlike 4.0 International"],"dc:subject":["Computer science"],"dc:title":["From Language Models to Moral Machines : Mechanism and Generalization"],"dc:type":["text","Electronic dissertations"]},"updated_at":"2026-11-18T18:59:34Z"}