BLiMP-IT

Dataset Summary

BLiMP-IT is a linguistically motivated benchmark for evaluating Italian language models through minimal pairs. Each example consists of a grammatical sentence paired with a minimally different ungrammatical counterpart that isolates a single morphosyntactic contrast. The benchmark is designed to evaluate whether language models assign higher probability to the grammatical sentence than to the ungrammatical one.

The benchmark is inspired by the English BLiMP while extending and adapting it to Italian morphosyntax. It combines phenomena from the original BLiMP with additional contrasts derived from Italian linguistic literature and the COnVERSA psycholinguistic assessment battery.

Rather than measuring downstream task performance, BLiMP-IT measures the acquisition of abstract morphosyntactic generalizations.

Supported Tasks

The dataset is intended for:

• Language model evaluation through sentence probability comparison
• Minimal-pair acceptability evaluation
• Psycholinguistically motivated syntactic evaluation
• Cross-model comparison
• Cross-linguistic comparison with BLiMP and related benchmarks

Typical evaluation consists of assigning higher probability (or lower perplexity) to the grammatical member of each minimal pair.

---

Dataset Structure

The repository contains one YAML configuration per linguistic phenomenon.

Macro-phenomena

The benchmark covers four major linguistic domains.

Macro-phenomenon	Number of phenomena
Agreement & Inflection	31
Verbal Class & Argument Structure	10
Pronouns	25
Non-local Dependencies	64

Overall the benchmark contains 130 phenomena.

Some phenomena originate from the original BLiMP benchmark, while others are adapted from the COnVERSA battery or specifically developed for Italian.

---

Phenomena

Agreement & Inflection

Original BLiMP

• DP agreement
  • blimp_it_original_agreement_and_inflection_dp

COnVERSA

• Determiner–noun agreement

  • conversa_a_agreement_dp
  • conversa_a_agreement_in_dp

• Subject–verb agreement

  • conversa_a_agreement_subject_verb_trans
  • conversa_a_agreement_subject_verb_unacc
  • conversa_a_agreement_subject_verb_unerg
  • conversa_a_agreement_subject_verb_unerg_trans
  • conversa_a_agreement_subject_verb_attraction
  • conversa_a_agreement_subject_verb_cumulative
  • conversa_a_agreement_verb_subject_unacc

• Predicate agreement

  • conversa_a_agreement_subject_nominal_predicate
  • conversa_a_agreement_subject_nominal_predicate_attraction

• Past participle agreement

  • conversa_a_agreement_past_particple_pre_v_cl
  • conversa_a_agreement_past_particple_unacc
  • conversa_a_past_particple_pre_v_cl
  • conversa_a_past_particple_unacc

• Psych verbs

  • conversa_a_agreement_psych_verbs_piacere
  • conversa_a_agreement_psych_verbs_preoccupare
  • conversa_a_psych_verbs_piacere
  • conversa_a_psych_verbs_preoccupare

---

Verbal Class & Argument Structure

COnVERSA

• Theta-role assignment

  • conversa_b_theta_roles
  • conversa_b_argument_structure_theta_roles
  • conversa_b_argument_structure_b_theta_roles

• Auxiliary selection

  • conversa_b_auxiliary_selection_transitives
  • conversa_b_auxiliary_selection_unaccusatives
  • conversa_b_auxiliary_selection_unergatives
  • conversa_b_auxiliary_selection_ditranstitives
  • conversa_b_auxiliary_selection_passive_active_diathesis

• Argument structure variants

  • conversa_b_argument_structure_b_auxiliary_selection_transitives
  • conversa_b_argument_structure_b_auxiliary_selection_unaccusatives
  • conversa_b_argument_structure_b_auxiliary_selection_unergatives
  • conversa_b_argument_structure_b_auxiliary_selection_ditranstitives
  • conversa_b_argument_structure_b_auxiliary_selection_passive_active_diathesis

---

Pronouns

COnVERSA

• Personal pronouns

  • conversa_c_pronouns_1_2_person_answers
  • conversa_c_pronouns_1_2_person_elicited_declaratives

• Clitics

  • conversa_c_pronouns_clitics_answering_with_acc_cl
  • conversa_c_pronouns_clitics_coordinating_with_acc_cl
  • conversa_c_pronouns_clitics_coordinating_with_dat_cl

• Reflexives

  • conversa_c_pronouns_reflexives_other
  • conversa_c_pronouns_reflexives_psych_verbs
  • conversa_c_pronouns_reflexives_unacc

---

Non-local Dependencies

COnVERSA

• Yes/no questions
  • conversa_d_answering_yes_no_questions
• Why questions
  • conversa_d_answering_why_questions
• WH-questions
  • conversa_d_answering_wh_adjunct_questions
  • conversa_d_answering_wh_argument_questions
  • conversa_d_answering_wh_argument_questions_number_disambiguation
  • conversa_d_answering_questions_with_rc_subj
  • conversa_d_answering_questions_with_rc_obj
  • conversa_d_question_formation_subject_position

Original BLiMP

• Relative clauses

  • subject relatives
  • object relatives
  • gap vs no-gap conditions

• WH extraction

  • root extraction
  • embedded extraction
  • two-level embedded extraction
  • clitic fillers
  • NP fillers
  • demonstrative fillers

• Across-the-board movement

  • affirmative
  • interrogative
  • relative clause
  • gap/no-gap manipulations

• Islands

  • adjunct island
  • complex NP island
  • sentential subject island
  • WH island
  • left branch island
  • coordinate structure constraint

• Parasitic gaps

  • adjunct island variants

---

Data Fields

Each dataset configuration consists of sentence pairs containing:

• sentence_good
• sentence_bad
• grammaticality label
• linguistic phenomenon identifier
• minimal-pair metadata

---

Data Creation

Source

BLiMP-IT combines:

• manually designed minimal pairs
• automatic lexical generation using grammar templates
• manually annotated Italian lexicon
• human validation

The lexicon was derived from a 2.4M-token corpus of Italian child-directed and naturalistic language, annotated with Universal POS tags, morphological features, and animacy information. Automatic generation produces lexical variants while preserving the targeted morphosyntactic contrast.

---

Dataset Characteristics

The benchmark follows two design principles:

1. Minimality of contrasts

   Each grammatical and ungrammatical sentence differs only in the manipulation responsible for (un)grammaticality.

2. Linguistically meaningful errors

   Ungrammatical sentences contain theoretically motivated morphosyntactic violations rather than random perturbations or word permutations.

---

Intended Uses

BLiMP-IT is intended for

• evaluating Italian language models
• probing syntactic generalization
• comparing language models
• comparing humans and language models
• psycholinguistic research
• computational linguistics research

---

Limitations

• The benchmark focuses exclusively on morphosyntax rather than semantics.
• Some generated sentences may be semantically unusual while remaining syntactically well formed.
• Human acceptability judgments currently exist only for subsets overlapping with COnVERSA.
• Automatic lexical generation is still expanding coverage across all phenomena.

---

Citation

Bressan et al. (2026) Language models assessment through linguistically motivated contrasts: A benchmark for Italian (BLiMP-IT). Proceedings of GLOW 47. DOI: 10.11576/glow-1242

@inproceedings{bressan2026blimpit,
  title={Language models assessment through linguistically motivated contrasts: A benchmark for Italian (BLiMP-IT)},
  author={Bressan, Veronica and Barbini, Matilde and Fusco, Achille and Neri, Sofia and Piccini Bianchessi, Maria Letizia and Rossi, Sarah and Chesi, Cristiano},
  booktitle={Proceedings of GLOW 47},
  year={2026}
}

---

License

CC BY-SA 4.0