A Hitchhiker’s Guide to Talking with Humans, Not LLMs: Visibility

nonasking · 2026-10 · 한국어판

This essay was written with heavy use of AI. Exactly how is disclosed at the end.

The same words can wear a person’s heart down or smooth it over.
This is a guide to talking with someone who has one.

There are many ways to tell someone who arrived late that you were hurt. You can say, “You were late and it upset me.” You can say, “You must be busy these days.” You can sit with a stiff face, or give no reply at all. What is being conveyed is one thing, ‘I am hurt’, and what differs is how much of it is in view. This essay calls that visibility. Every example not taken from the literature belongs to this one scene: someone who waited, and someone who was late. In these examples, what is actually said is in double quotes, and the message meant is in single quotes.

Explicit and Implicit

The Standard Korean Language Dictionary glosses the explicit as “showing something plainly” and the implicit as “showing one’s meaning tacitly, without expressing it directly in words or action.” The implicit, too, shows meaning. It differs in how.

Politeness theory divided the how into five. When a person is about to do something that may touch the other’s face, they say it baldly, say it softened with positive politeness, say it softened with negative politeness, say it off record, or do not say it at all. The theory calls the size of the act its weightiness. Weightiness is the sum of the distance between the two, the difference in power, and the burden that act carries in that culture, and the heavier the act, the further down the list the choice falls. Under the five ways hang concrete means. Under speaking off record there are fifteen strategies, such as giving hints, over-generalizing, and displacing the hearer (Brown & Levinson 1987).

Research on requests uses the same frame. On a scale of directness it places three levels, direct, conventionally indirect, and non-conventionally indirect, and lays nine specific types along them (Blum-Kulka, House & Kasper 1989).

This essay also divides in two. Explicit and implicit are the scale. Silence, face and body, symbolic action, generalized remarks, and implicating remarks are the means.

The scale has four parts. There is precedent. One analysis holds that a message is made of four elements, “I am saying this to you in this situation” (Haley 1959). Later work defined a blurred message as one that obscures at least one of the four, and scored messages along these four as dimensions (Bavelas, Black, Bryson & Mullett 1986). Of this essay’s four, the proposition and the addressee come from there; the receipt and the demand for a response are this essay’s own.

• Proposition: is the content to be conveyed written on the surface?
• Addressee: is it shown on the surface who is being addressed?
• Receipt: does the fact of having conveyed it remain between the two?
• Demand for a response: must the one who received it answer?

When all four are on the surface, that is explicit. “You were late and it upset me.” The ‘I am hurt’ being conveyed is written in the sentence as is.

Politeness theory sets the payoffs of speaking on record and off record side by side. The one who speaks on record gets credit for honesty and candor, avoids the suspicion of manipulation, and avoids the risk of being misunderstood. The one who speaks off record gets credit for tact and for not being coercive, is not caught by a face-damaging interpretation, and gives the hearer a chance to show care (Brown & Levinson 1987).

The explicit demands an answer. The one who receives it cannot pretend not to have heard. When someone who has received a response-seeking utterance fails to produce a recognizable response, that conduct becomes something that calls for an account (Pomerantz 1984).

Read through these four parts, the five means and the explicit look like this.

The more parts on the surface, the longer the visibility. Only the two ends have a fixed order. Silence has none of the four and the explicit has all four. Among the means in between there is no fixed order.

What the table measures is what is written on the surface. It does not measure whom the speaker actually aimed at, or what they actually meant.

The means overlap. One can make an implicating remark with a stiff face, or post a generalized remark on a profile. Research that sorted the methods of signaling also noted that one signal can combine several methods (Clark 1996). There is also a theory that places words and non-words in a single space. From meaning to showing, from determinate content to content that cannot be determined, two continua stand at right angles, and what one intends to convey can lie anywhere in that space (Sperber & Wilson 2015).

The previous essay (in Korean) dealt with receiving words. If receiving words adjusts how much of the other’s words one takes on, sending words adjusts how much one is committed to one’s own.

Silence

The one who waited does not reply to the latecomer’s message. What is being conveyed is ‘I am hurt.’ Silence has none of the four. Even so, something is conveyed. An old axiom of communication theory is that one cannot not communicate. Whether one speaks or not, it has message value (Watzlawick, Beavin & Jackson 1967).

How long a silence must be before it carries meaning has been measured. In a study of 195 responses in telephone conversation, only when the gap between turns was 700 ms or more was the proportion of responses like refusal or disagreement clearly higher than that of acceptance or agreement. Looking at the type of response alone, the timing of the two did not divide systematically at shorter gaps (Kendrick & Torreira 2015).

The listener’s ear reacts at a similar point. When the gap after a request or an assessment was manipulated to 0 ms, 600 ms, and 1200 ms, the longer the gap, the weaker the willingness to comply sounded and the weaker the agreement sounded (Roberts, Francis & Morgan 2006). In an experiment that sliced the gap after requests more finely, ratings dropped noticeably at 600 ms and differed significantly between 700 and 800 ms (Roberts & Francis 2013).

What silence is read as depends on what came before it. A long silence after a request or an assessment reads toward refusal and disagreement. Silence after an assertion can be read as acceptance (Farkas & Bruce 2010).

What a speaker can do when no response comes has been sorted into three. Give up on getting one. Infer the other’s response and leave it unspoken. Keep pursuing a spoken response (Pomerantz 1984). On the second path, the meaning of the silence is filled in by the one who received it. The one who was silent can say “I didn’t say anything,” and that is true.

Face and Body

The one who waited sits with a stiff face. What is being conveyed is ‘I am hurt.’ That face may be seen by the person sitting across. No words were addressed, so the addressee is only half in view. Even if it was seen, what was shown does not remain, so the receipt is half too. There is no proposition and no demand for a response.

The expressions a person gives off divide in two. One is the expression given: words or their substitutes, used admittedly and for the purpose of conveying information. The other is the expression given off: a wide range of conduct that others can treat as symptomatic of the person. The observer uses the side thought to be hard to control as a check on whether what the controllable side conveys is true (Goffman 1959).

Among expressions given off, the parts of the body differ. The study that named the escape of information one meant to hide leakage wrote that the face is likely to be the major nonverbal liar, and that the legs and feet are a primary source of leakage (Ekman & Friesen 1969).

People communicate in two modes, digital and analogic. Digital language, which corresponds to words, has a complex and powerful logical syntax but lacks adequate semantics in the field of relationship. Analogic language, which corresponds to facial expression, gesture, and tone, has that semantics but no syntax for defining the nature of a relationship unambiguously (Watzlawick, Beavin & Jackson 1967).

So a stiff face conveys something and yet cannot tell ‘I am hurt’ from ‘I am tired’ from ‘I am not interested.’ The telling is done by the one who looks.

Symbolic Action

That evening, the one who waited changes their messenger profile picture to a photo of a clock. What is being conveyed is ‘I am hurt.’ Such an action is open to everyone and has no fixed addressee.

Before decoding the meaning, the receiving side must first decide whether this was meant for them. Even having decided it was, and decoded it, there is no way to confirm. The action remains; the meaning does not. Ask, and “I just changed it” may come back. It is the reverse of the face. The face may have been conveyed but leaves nothing behind; the picture remains but there is no knowing whether it was conveyed. For both, the receipt is half.

Implicatures come strong and weak. When what the speaker wants to call to mind is narrowed to one thing, the implicature is strong. The wider the range of possible alternatives, the weaker the implicature, and the more the hearer bears some share of responsibility for the premises they themselves supplied (Wilson & Sperber 2012). What a symbolic action carries is a bundle of weak implicatures.

That expressions given off are used to check expressions given lends this means its use, because the side thought to be hard to control can also be staged. So an information game of concealment, discovery, false revelation, and rediscovery can go on without end (Goffman 1959).

Generalized Remarks

“I really can’t stand people who don’t keep time.” What is being conveyed is ‘I am hurt by you,’ and ‘by you’ is not in the sentence. There is a proposition and there is a receipt. There is no addressee. From the surface alone, one cannot tell whether it is really a general remark or aimed at someone.

Politeness theory lists this means separately among the fifteen off-record strategies. The thirteenth is over-generalizing. State a rule, and whom the act is aimed at hides in a blur. The example is “Mature people sometimes help do the dishes.” Whether the general rule applies to oneself in this case becomes the hearer’s choice. Proverbs do the same work (Brown & Levinson 1987).

The fourteenth strategy is displacing the hearer. Instead of speaking directly to the person actually aimed at, one addresses someone else whose face the words will not damage, and hopes the real target will overhear (Brown & Levinson 1987). In front of the latecomer, the one who waited says to another friend who came along, “You’re always on time.” What is being conveyed is ‘you were late, and I am hurt.’

What one ends up saying to take back a remark made this way is “I wasn’t talking about you.” That one dislikes it cannot be taken back. What can be taken back is the addressee.

Where there are several possible addressees, each decides whether the words were theirs. Research on organizational communication called the deliberate use of ambiguity to accomplish one’s goals strategic ambiguity. It lets several perspectives stand together, preserves privileged positions, and can be denied. Ambiguous communication is marked by its assertorial lightness and is therefore more easily denied than its less equivocal counterpart (Eisenberg 1984). The ambiguity this research means covers blurred speech broadly.

Where there is only one possible addressee, a generalized remark works differently. Whose words they are, arriving in the form of a general rule, is settled without needing to be settled.

Implicating Remarks

“You must be busy these days.” Addressed directly to the other, so there is an addressee. The ‘I am hurt’ being conveyed is not in the sentence; the hearer has to infer it back. What was said remains; what was meant does not.

Where this parts from the earlier means is in what can be taken back. The face is taken back with “I’m just tired,” the profile with “I didn’t post it for you,” the generalized remark with “I wasn’t talking about you.” All three take back the very fact of having tried to convey something to you. The implicating remark cannot take back that words were addressed; it can only take back the content. “That’s not what I meant.” One study defines insinuation as communication that is partly overt and partly covert (Witek 2024).

Conversational implicatures can be cancelled. They can be cancelled explicitly by adding a clause that states or implies the speaker has opted out, and they can be cancelled contextually (Grice 1975; 1978). What one ends up saying to take back an implicating remark is “I was just asking how you’ve been.” What can be taken back is the proposition. The addressee stays.

One theory explaining indirect speech names three things. An indirect request can be plausibly denied. Language does the work of conveying information and the work of negotiating the type of relationship at once. And people perceive language as a digital medium, so that a single sentence can create common knowledge. The example is a driver pulled over: “Gee, officer, is there some way we could take care of the ticket here?” Indirect speech provides only individual knowledge; direct speech provides common knowledge (Pinker, Nowak & Lee 2008). Both may know, and yet it never becomes the state in which both know that both know.

Philosophy treats this under the name insinuation. Insinuation is implicature that leaves room for denial about risky attitudes and contents, and conveys the content without entering it on the conversational record (Camp 2018). Deniability is also defined as an epistemic matter. If the speaker says that was not what they meant, the hearer no longer knows that it was (Dinges & Zakkou 2023).

Experiments measure what implicature relieves and what it does not. Asked whom they would trust between a speaker who asserted content later shown to be false and one who implicated it, 30 of 41 participants (73%) chose the one who implicated. On the question of whom to hand something close to punishment, there was no significant difference (Mazzarella, Reinecke, Noveck & Mercier 2018). In an experiment on promises, people saw speakers as committed not only to what they said but to what they meant (Bonalumi, Scott-Phillips, Tacha & Heintz 2020). In an experiment on denials, how far the speaker was held responsible depended not on the denial itself but on whether there had been an incentive to mislead (Bonalumi, Bumin, Scott-Phillips & Heintz 2023). Whether something is judged a lie is closely bound to how far the person is seen as having committed themselves (Reins & Wiegmann 2021).

Implicature is not the only way to blur a proposition. An ambiguous remark splits into two or more readings. “Well, you took your time” reads as envy of someone’s ease or as sarcasm. What is being conveyed is ‘you were late and unbothered, and I am hurt.’ A vague remark has no boundary. “Try to come a bit earlier next time” does not settle how early counts as keeping it. What is being conveyed is ‘be on time next time.’ A general remark has low resolution. “Today was a bit off” is true and carries nothing about what was off. What is being conveyed is ‘you were late and I was hurt.’

Semantics separates these three. Ambiguity is uncertainty in the mapping between expression and meaning; vagueness is when the meaning of a word itself has borderline cases (Kennedy 2011). Being underspecified differs from being ambiguous. The English ‘aunt’ does not settle whether it is one’s mother’s sister or one’s father’s, but it is not a word with two meanings (Sennet 2021). A study of ostensible invitations listed, among seven features of such invitations, leaving the arrangements vague (Isaacs & Clark 1990). What the receiving side has to do also differs by path. Implicature must be inferred, ambiguity must be chosen between, vagueness must have a line drawn, and low resolution must be filled in.

The Exit

The demand for a response can be detached from the other parts. “I was a bit hurt that you were late. You don’t have to answer.” What is being conveyed is ‘I was hurt’ and ‘you don’t have to answer,’ and both are written in the sentence. There is a proposition, an addressee, and a receipt. The exit, that no answer is required, is attached inside the sentence.

Politeness theory places this under negative politeness. Not coercing the hearer’s response, explicitly giving the option not to do the act, is one way of redressing the hearer’s face (Brown & Levinson 1987). Earlier work set down three rules of politeness. Don’t impose. Give options. Make the other feel good (Lakoff 1973).

Where the Weight Rests

The table below is not the literature’s but the frame this essay sets out to look through; what follows it is that frame propped up with the literature. What one intends to convey has weight. Whether sent in the open or sent blurred, the weight is not set down. Its kind changes.

Being explicit, the sender is committed to their words. They are bound to what they said, and taking it back costs something. The receiver has nothing to decode and must answer. Sending it blurred, the sender’s commitment shrinks, and whether it landed stays unknown. The receiver need not answer and must decode.

What disappears when one sends it blurred is not the weight but the receipt. People see a speaker as bound to what they meant as well (Bonalumi, Scott-Phillips, Tacha & Heintz 2020). It is already entered in the receiver’s ledger; it is missing only from the record between the two.

Indirectness is said to carry two large payoffs: rapport and self-defense (Tannen 1986). Rapport is the satisfaction of being understood without explaining oneself and of getting what one wants without asking for it. Defense is being able to step back and save face when one’s words are not received well (Tannen 1981). When the one who waited tries to convey ‘I am hurt’ with “You must be busy these days,” and the latecomer understands and apologizes, the one who waited is understood without having said it. If the latecomer does not understand, or bristles, the one who waited can step back with “I was just asking how you’ve been.”

The explicit end has its price too. Words are precise in carrying content and clumsy in carrying relationship (Watzlawick, Beavin & Jackson 1967). So words written on the explicit side are easily called cold. Move down toward the implicit side to carry warmth, and what the receiver must decode grows. The path that remains is to transcribe the relationship into a proposition. Hurt can be shown in a face or written as the sentence “I was hurt.”

Weight is not the only reason to choose the implicit side.

There are times one does not know. The silence of someone who does not yet know whether they are hurt or tired looks, from outside, the same as the silence of someone holding a proposition and not sending it. A vague remark can be the more accurate one. Someone who does not know exactly how late they will be, trying to convey ‘I don’t know exactly, but about thirty minutes,’ may give the listener more information by saying “I’ll be about thirty minutes late” than by naming a precise interval (Égré, Spector, Mortier & Verheyen 2023).

There are times the other cannot receive. The day the latecomer walks in with an exhausted face is one. The explicit demands an answer. On such a day, “You must be busy these days” tries to convey ‘I am hurt’ while leaving someone not ready to answer an exit to pretend not to notice.

There are times the relationship comes before the proposition. Relationship rides better on face and body. The latecomer walks in, and the one who waited smiles without a word and makes room. What is being conveyed is ‘it’s fine, I’m glad you’re here.’

There are times the proposition is not the speaker’s alone. If the reason for being late lies with family, that reason is not the latecomer’s to bring up. “Something came up at home” is true and only low in resolution. What is being conveyed goes as far as ‘there was a reason worth being late for.’ There is a theory that private information can become jointly owned, and that those who jointly own it are bound by rules about whether to spread it further (Petronio 2002). Borrowing that frame, what is missing here was not cut to lighten one’s commitment but left out because it belongs to someone outside the sentence. Saying in the same place “That’s not my story to tell” makes the concealing itself explicit. What is being conveyed is ‘there is a reason, and I cannot tell it.’ It is the path of declaring that one is opting out of the maxim. “I cannot say more; my lips are sealed” (Grice 1975).

The receiving side has a share too. Weighing the grounds when deciding whether a generalized remark is theirs, and not nailing a silence to a single meaning, is that share.

Visibility does not measure honesty. There is silence out of not knowing, and there is explicitness that is false. What it measures is where the weight has been placed.

Who is carrying, right now, the weight that could have been lifted?

References

• National Institute of Korean Language. Standard Korean Language Dictionary. Entries for 명시2(明示) and 묵시1(默示).
• Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge University Press.
• Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-cultural pragmatics: Requests and apologies. Ablex.
• Haley, J. (1959). An interactional description of schizophrenia. Psychiatry, 22(4), 321-332. doi:10.1080/00332747.1959.11023187
• Bavelas, J. B., Black, A., Bryson, L., & Mullett, J. (1986). A situational explanation of discourse: The case of political disqualification. Working Papers of the Linguistics Circle of the University of Victoria, 5(1), 33-44.
• Bavelas, J. B., Black, A., Chovil, N., & Mullett, J. (1990). Equivocal communication. Sage.
• Pomerantz, A. (1984). Pursuing a response. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action (pp. 152-164). Cambridge University Press.
• Clark, H. H. (1996). Using language. Cambridge University Press.
• Sperber, D., & Wilson, D. (2015). Beyond speaker’s meaning. Croatian Journal of Philosophy, 15(44), 117-149.
• Watzlawick, P., Beavin, J. H., & Jackson, D. D. (1967). Pragmatics of human communication: A study of interactional patterns, pathologies, and paradoxes. W. W. Norton.
• Kendrick, K. H., & Torreira, F. (2015). The timing and construction of preference: A quantitative study. Discourse Processes, 52(4), 255-289. doi:10.1080/0163853X.2014.955997
• Roberts, F., Francis, A. L., & Morgan, M. (2006). The interaction of inter-turn silence with prosodic cues in listener perceptions of “trouble” in conversation. Speech Communication, 48(9), 1079-1093. doi:10.1016/j.specom.2006.02.001
• Roberts, F., & Francis, A. L. (2013). Identifying a temporal threshold of tolerance for silent gaps after requests. The Journal of the Acoustical Society of America, 133(6), EL471-EL477. doi:10.1121/1.4802900
• Farkas, D. F., & Bruce, K. B. (2010). On reacting to assertions and polar questions. Journal of Semantics, 27(1), 81-118. doi:10.1093/jos/ffp010
• Goffman, E. (1959). The presentation of self in everyday life. Doubleday Anchor.
• Ekman, P., & Friesen, W. V. (1969). Nonverbal leakage and clues to deception. Psychiatry, 32(1), 88-106. doi:10.1080/00332747.1969.11023575
• Wilson, D., & Sperber, D. (2012). Meaning and relevance. Cambridge University Press.
• Eisenberg, E. M. (1984). Ambiguity as strategy in organizational communication. Communication Monographs, 51(3), 227-242. doi:10.1080/03637758409390197
• Witek, M. (2024). Gricean insinuation and the fake one-way mirror effect. Synthese, 204, 96. doi:10.1007/s11229-024-04699-5
• Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics 3: Speech acts (pp. 41-58). Academic Press.
• Grice, H. P. (1978). Further notes on logic and conversation. In P. Cole (Ed.), Syntax and semantics 9: Pragmatics (pp. 113-127). Academic Press.
• Pinker, S., Nowak, M. A., & Lee, J. J. (2008). The logic of indirect speech. Proceedings of the National Academy of Sciences, 105(3), 833-838. doi:10.1073/pnas.0707192105
• Camp, E. (2018). Insinuation, common ground, and the conversational record. In D. Fogal, D. W. Harris, & M. Moss (Eds.), New work on speech acts (pp. 40-66). Oxford University Press. doi:10.1093/oso/9780198738831.003.0002
• Dinges, A., & Zakkou, J. (2023). On deniability. Mind, 132(526), 372-401. doi:10.1093/mind/fzac056
• Mazzarella, D., Reinecke, R., Noveck, I., & Mercier, H. (2018). Saying, presupposing and implicating: How pragmatics modulates commitment. Journal of Pragmatics, 133, 15-27. doi:10.1016/j.pragma.2018.05.009
• Bonalumi, F., Scott-Phillips, T., Tacha, J., & Heintz, C. (2020). Commitment and communication: Are we committed to what we mean, or what we say? Language and Cognition, 12(2), 360-384. doi:10.1017/langcog.2020.2
• Bonalumi, F., Bumin, F. B., Scott-Phillips, T., & Heintz, C. (2023). Communication and deniability: Moral and epistemic reactions to denials. Frontiers in Psychology, 13, 1073213. doi:10.3389/fpsyg.2022.1073213
• Reins, L. M., & Wiegmann, A. (2021). Is lying bound to commitment? Empirically investigating deceptive presuppositions, implicatures, and actions. Cognitive Science, 45(2), e12936. doi:10.1111/cogs.12936
• Kennedy, C. (2011). Ambiguity and vagueness: An overview. In C. Maienborn, K. von Heusinger, & P. Portner (Eds.), Semantics: An international handbook of natural language meaning (Vol. 1, pp. 507-534). De Gruyter Mouton. doi:10.1515/9783110226614.507
• Sennet, A. (2021). Ambiguity. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/ambiguity/
• Isaacs, E. A., & Clark, H. H. (1990). Ostensible invitations. Language in Society, 19(4), 493-509. doi:10.1017/S0047404500014780
• Lakoff, R. (1973). The logic of politeness; or, minding your p’s and q’s. In C. Corum, T. C. Smith-Stark, & A. Weiser (Eds.), Papers from the Ninth Regional Meeting of the Chicago Linguistic Society (pp. 292-305). Chicago Linguistic Society.
• Tannen, D. (1986). That’s not what I meant! How conversational style makes or breaks your relations with others. William Morrow.
• Tannen, D. (1981). Indirectness in discourse: Ethnicity as conversational style. Discourse Processes, 4(3), 221-238.
• Égré, P., Spector, B., Mortier, A., & Verheyen, S. (2023). On the optimality of vagueness: “around”, “between” and the Gricean maxims. Linguistics and Philosophy, 46(5), 1075-1130. doi:10.1007/s10988-022-09379-6
• Petronio, S. (2002). Boundaries of privacy: Dialectics of disclosure. State University of New York Press.

This essay was written out of the records of conversations with an AI, through dialectical engagement between me and the machine, typed with the AI’s borrowed hands.