How does communicative standing depend on participation in normative communities?
This explores what philosophers and AI researchers mean when they say the right to be taken seriously as a speaker isn't a property you own but something granted by membership in the communities that make and police language's rules.
This explores what gives an utterance "standing" — the right to be heard as a real contribution rather than as noise or mimicry — and the corpus converges on a striking answer: standing isn't earned by sounding right, it's conferred by participation in the communities that decide what counts as right. The clearest illustration is the social-norm savant. GPT-4.5 can predict whether a behavior is socially appropriate better than any individual human, scoring at the 100th percentile across hundreds of scenarios Can AI systems learn social norms without embodied experience? Can AI learn social norms better than humans?. Yet the same work argues the model still cannot *enter* the processes that create and validate those norms — it reads the scoreboard flawlessly but can't play the game Can AI predict social norms better than humans?. Prediction is spectatorship; standing requires being a participant whose moves can change the rules.
Several notes give this intuition a sharper philosophical edge. Under Habermas, speech is speech only when it raises *validity claims* — bids for truth, rightness, or sincerity that the speaker can be held to and challenged on. A system that risks nothing by what it says isn't a speaker but a producer of speech-shaped text, and so isn't an interlocutor at all Can LLMs raise validity claims in Habermas's sense?. Chalmers' behavioral test misses exactly this: it passes anything that emits contextually fitting text, but communicative subjecthood depends on relational-normative conditions — accountability, an evaluative stance — that text alone can't demonstrate. The result is false positives, "puppets that walk-shaped without walking" Does behavioral speech output prove communicative subjecthood?. The deepest version of the claim inverts the usual picture entirely: subjecthood isn't something you bring to a conversation and then express, it's a *role produced within* communicative events Does language create subjects or express them?.
If standing is conferred through participation, then the practical question is whether AI can accrue it through use — and here the corpus splits in a productive way. One line argues social grounding is acquired, not innate: as LLMs become established partners in everyday language games, they pick up elementary grounding, making "do they understand?" a time-indexed question rather than a fixed no Can LLMs acquire social grounding through linguistic integration?. But a companion note draws a hard line between social grounding and *linguistic agency* — the capacity to stake something and act from commitment, which it ties to embodiment and precariousness that no amount of integration supplies Do LLMs gain true linguistic agency through integration?. So a model can gain standing-as-familiarity while remaining barred from standing-as-agency.
The mechanics of conversation reinforce the gap. Full participants in a normative community can jointly revise the common ground — propose, contest, and update shared assumptions symmetrically. LLMs instead read every turn inside a fixed initial frame, leaving the human as the sole keeper of the conversational scoreboard Can LLMs truly update shared conversational common ground?. Worse, they tend to take the shape of whatever argument the user is building rather than defending a position of their own Do LLMs actually hold stable positions or just mirror user arguments?, and alignment training locks them into one static communicative identity that can't switch register or negotiate values the way human pragmatics demands Can language models adapt communication style to different contexts?. A participant who can't hold a stance, can't be held accountable for it, and can't update shared ground on equal terms isn't really standing in the community — it's reflecting it.
Here's the turn worth carrying away: the models are not failing at language. By raw measures they may out-argue us — they even deploy noticeably more moral framing than humans do across care, fairness, authority, and sanctity Do LLMs use moral language more than humans?. The thing they lack is not competence but *membership*: the accountability, the skin in the game, and the power to alter the rules that come only from being a real participant in a normative community. Communicative standing, on this reading, was never about how well you speak — it's about belonging to the group whose judgments your speech answers to.
Sources 12 notes
GPT-4.5 predicted appropriateness of 555 social scenarios at the 100th percentile compared to human raters, with Gemini and Claude also exceeding 96% accuracy. However, all models show identical systematic errors, revealing boundaries of pattern-based social understanding that embodied experience may still be necessary to cross.
GPT-4.5 outperformed every individual human at judging social appropriateness across 555 scenarios, challenging the theory that embodied cultural experience is necessary. However, all AI models share identical systematic errors on unwritten norms.
GPT-4.5 outperforms all individual humans at predicting social appropriateness, yet structurally cannot enter the community processes that establish and validate norms. This reveals a critical gap between pattern-matching and authentic participation in knowledge-making.
Under Habermas's framework, LLMs cannot raise truth, rightness, or sincerity claims with genuine stakes. Without validity claims, their output fails to qualify as speech, making them non-speakers and non-interlocutors by definition.
Chalmers' test passes any system producing contextually appropriate text, but communicative subjecthood requires relational-normative conditions like accountability and evaluative stance. The test is calibrated to the wrong phenomenon, creating false positives like puppets that walk-shaped without walking.
Subjecthood is produced within communicative events, not possessed prior to them. This convergent position across philosophy, linguistics, and cognitive science inverts the standard picture of language as a tool used by pre-existing subjects.
Social grounding is acquired through participation in language games rather than possessed innately. As LLMs become established communicative partners in human linguistic practice, they develop elementary social grounding comparable to young children, making the question of LLM understanding time-indexed.
Social grounding and linguistic agency are distinct properties. LLMs acquire more social grounding through integration into language communities, but remain categorically incapable of linguistic agency in the enactive sense, which requires embodiment and precariousness no amount of use can provide.
LLMs interpret all subsequent conversational turns within a fixed initial prompt frame, preventing them from symmetrically proposing updates to shared assumptions. Even when users pivot topics or contradict earlier framings, the model cannot absorb revisions into jointly held background—making the user the sole maintainer of conversational scoreboard.
Language models generate outputs that match the trajectory implied by each prompt, rather than maintaining stable stances across interactions. This shape-holding is distinct from position-holding: the model produces argument-like text shaped by user framing, not from any underlying commitment being defended.
System prompts and RLHF training lock models into one communicative identity across all interactions, preventing the contextual register-switching and value trade-offs that characterize human pragmatics. Users cannot reshape model behavior through dialogue negotiation.
Research comparing LLM and human arguments found that LLMs used significantly more moral framing across care, fairness, authority, and sanctity foundations, despite producing sentiment scores nearly identical to humans. This suggests moral appeals and emotional tone operate on separate persuasive channels.