arXiv subjects

Manuel Pita

Publications and source records attributed to Manuel Pita.

4 recordsLinked to original sources

arXiv2026

Trusting sovereign language models as scientific instruments: evidence from Portugal's AMALIA

National language models are becoming publicly funded epistemic infrastructure. Public ownership, linguistic specialization, and open weights create a presumption of trustworthiness. Such an instrument, built by and for a language community, looks like the natural choice for measuring what that community says and values. Whether such a model validly measures anything is untested at release. The evaluation of LLMs as measurement instruments is typically task-specific and stops at agreement with human coders. Agreement cannot distinguish an LLM instrument that measures a construct from one that reaches matching codes through surface correlates. We audit the presumption on a favourable case: AMALIA, Portugal's publicly funded 9B model, coding the moral foundation of authority in European Portuguese. The \textit{recovery gap} operationalizes the audit: decompose the codebook into its theory-defined clauses, recombine them through the theory's explicit rule, and measure how much of the original prompt's performance the stated theory reproduces. In a pre-registered, out-of-sample study on a transcreated (English to European Portuguese) corpus, AMALIA agrees with trained coders within six points of open models eight to thirteen times its size. Yet, the recovery gap shows that only about half of coding performance on authority can be attributed to the theory. A larger multilingual LLM closes the recovery gap on the same corpus, suggesting the shortfall lies in the annotator model, not the corpus or its translation. Sovereignty earns operational and performance trust; epistemic trust requires calibration -- and the audit method is inexpensive, and portable across models, languages and tasks.

cs.CL↗

arXiv2026

Correct codes for the wrong reasons? validating LLMs as measurement instruments for theoretical constructs

When a large language model (LLM) codes a construct in text as a human annotator would, that agreement makes the LLM a reliable coder. Yet reliability leaves construct validity untouched. The instrument may be theory-naive, reaching the code through a correlate that meets none of the demands the construct's theory makes, and no current method tells that apart from genuine measurement. We propose grain calibration as a method that closes the gap. It decomposes a construct into clause-level components, tests each against the text with extractive evidence, and combines the results through an explicit, theory-derived rule. Because the rule is stated rather than lodged in one opaque pass, its structure is evidence about the process rather than the output. It shows which components settled a code, and, when the code is wrong, whether a component was missed or an adjacent construct mistaken for it. Validation shifts from scoring an instrument's outputs against an annotator to showing that the instrument runs on the construct its theory specifies.

cs.CL↗

arXiv2026

How local rules generate emergent structure in cellular automata

Cellular automata generate spatially extended, temporally persistent emergent structures from local update rules. No general method derives the mechanisms of that generation from the rule itself; existing tools reconstruct structure from observed dynamics. This paper shows that the look-up table contains a readable causal architecture and introduces a forward model to extract it. The key observation in elementary cellular automata (ECA) is that adjacent cells share input positions, so the prime implicants of neighbouring transitions overlap. That overlap can couple the transitions causally or leave them independent. We formalize each pairwise interaction as a tile. A finite-state, tiling transducer, $\mathcal{T}$, composes tiles across the CA lattice, tracking how coupling and independence propagate from one cell pair to the next. Structural properties of $\mathcal{T}$ are used to classify ECA rules that can sustain regions of causal independence across space and time. We find that, in the 88 ECA equivalence classes, the number of local configurations at which coupling is structurally impossible -- computable from the look-up table -- predicts the prevalence of dynamically decoupled regions with Spearman $ρ= 0.89$ ($p < 10^{-31}$). The look-up table encodes not just what a rule computes but where it distributes causal coupling across the lattice; the framework reads that distribution forward, from local logical redundancy to emergent mesoscopic organization.

nlin.CG↗

arXiv2023

Use of explicit replies as coordination mechanisms in online student debate

People in conversation entrain their linguistic behaviours through spontaneous alignment mechanisms [7] - both in face-to-face and computer-mediated communication (CMC) [8]. In CMC, one of the mechanisms through which linguistic entrainment happens is through explicit replies. Indeed, the use of explicit replies influences the structure of conversations, favouring the formation of reply-trees typically delineated by topic shifts [5]. The interpersonal coordination mechanisms realized by how actors address each other have been studied using a probabilistic framework proposed by David Gibson [2,3]. Other recent approaches use computational methods and information theory to quantify changes in text. We explore coordination mechanisms concerned with some of the roles utterances play in dialogues - specifically in explicit replies. We identify these roles by finding community structure in the conversation's vocabulary using a non-parametric, hierarchical topic model. Some conversations may always stay on the ground, remaining at the level of general introductory chatter. Some others may develop a specific sub-topic in significant depth and detail. Even others may jump between general chatter, out-of-topic remarks and people agreeing or disagreeing without further elaboration.

cs.CL↗