Saying “I do not know” while holding responsibility

Epigraph

follow your reason

Thomas Henry Huxley, “Agnosticism” (1889), Collected Essays, Volume V, p. 246; a brief fragment. The complete passage concerns intellectual questions and the reach of available reasoning. It does not require rejecting others' assistance or specialist knowledge.

What this article is about

When people expect an answer, admitting that you do not know can feel like abandoning your role. You arrange a shared meeting, explain how the work will proceed, or accept a modest commitment—and then face a question you cannot yet answer precisely. You might cover the gap with a confident tone, pass the question to someone else, or offer an indefinite wait and see. Another possibility is to identify what remains unknown, what has been confirmed, and what answer you can properly give now. That requires distinguishing the boundary of knowledge from the boundary of responsibility. They do not always fall in the same place.

We will explore this through an imagined small voluntary project. Lida is coordinating a drawing exhibition with acquaintances. The authors invented the scene; it is not a research case. John Locke's distinction between knowledge and judgment provides a narrow philosophical reference. The rest is our reasoning about keeping commitments intelligible without complete certainty. It does not replace specialist requirements, promise to prevent mistakes, or oblige anyone to explain private matters under pressure.

I. A role creates an expectation of an answer, not knowledge itself

Lida became coordinator because she agreed to gather proposals and keep participants in contact. At the first meeting, someone asks whether there will be room for every drawing. She knows the dimensions of some sheets but has not seen the final selection. The question sounds straightforward, and she wants to give a straightforward answer. Yet a confident yes would claim more than the available information supports. Becoming coordinator did not add a single measurement to that information. It created a responsibility to address the question, not an ability to know every outcome in advance.

This distinction is easy to lose in a small shared undertaking. Nobody demands omniscience from Lida, but having one contact person is convenient. Participants may assume that a question has already been checked because she speaks for the project. Her answer therefore means more than her intention to reassure them. Its meaning also depends on the actions they reasonably associate with it. If her confirmation leads them to stop considering another way to display the drawings, unsupported confidence has begun to shape their decisions. A benevolent intention does not make that effect neutral.

Responsibility does not necessarily belong entirely to one person. Lida collects information about formats, another participant selects drawings, and a third arranges access to the room. Making that division explicit before something goes wrong can be useful. Otherwise, everyone may believe that somebody else performed the final check. An admission of uncertainty can reveal an unclaimed task: nobody has yet compared the final dimensions with the available surface. That concerns how the shared work is organized. It does not establish that the coordinator is unsuitable for her role.

Nor does an appointment create an unlimited duty to answer for everything. Lida may have agreed only to relay information while others quietly added responsibility for choosing every condition. The agreement then needs clarification. She can say that a question exceeds the participation she accepted and discuss who is willing to take it on. This differs from pretending not to know in order to evade a promise. The distinction depends on the actual commitment, not on how persuasively someone refuses. Recovering the role requires looking at the agreement rather than at another person's expectation of endless availability.

II. Uncertainty becomes clearer when it has an object

A general I do not know leaves participants with too many possible interpretations. Lida may lack the drawing dimensions, not understand the hanging method, doubt her own choice, or have no information about access times. Those gaps require different actions. Naming the object narrows the answer: the final selection is not yet settled, so she cannot confirm that all sheets will fit at once. This does not resolve the question, but it identifies which conclusion cannot currently count as established. It preserves what is known rather than declaring the entire project unknown.

Sometimes the object of uncertainty emerges only after the question is clarified. One participant asks about space with the distance between drawings in mind; Lida thinks about the number of walls. Both can discuss the exhibition in good faith while answering different questions. Another wall measurement will not necessarily resolve that difference. They first need to understand which display arrangement they want to check. Clarifying meaning is not inherently a delaying tactic. It makes possible an answer that addresses the other person's question rather than merely containing an accurate number.

Confirmation and assumption can be kept separate. Lida has seen one wall and knows its length. She assumes that the adjacent wall will be available but has not received an answer. She likes the idea of spacing the drawings more freely. These are three different grounds. A conversation can blend them into the room is suitable, although one concerns observation, another expected access, and the third preference. Admitting uncertainty does not require dismissing all three. It requires avoiding presenting an assumption as confirmation or a preference as a property equally important to everyone.

Every gap need not be disclosed to every participant in full detail. Information may concern another person's privacy or be irrelevant to the shared action. Lida does not have to reveal someone's personal reason for being absent to say that their participation has not been confirmed. Precision is not the same as complete openness. She can communicate the limit of an available conclusion without passing on a private explanation. If the decision genuinely requires further information, how to obtain it appropriately remains a separate question. Wanting to finish the discussion does not remove another person's right to decline.

III. Lack of certainty does not make every judgment arbitrary

In Book IV of his Essay Concerning Human Understanding, Locke distinguishes certain knowledge from judgment where certainty is lacking. Chapters XIV–XVI examine grounds for probable assent, including experience, testimony, and scrutiny. This is historical philosophy within its own religious framework. Our narrow reference is the distinction between acting without certainty and presenting a supposition as knowledge. It neither establishes a modern coordinator's standard nor guarantees a correct conclusion.

Lida may have enough grounds for a provisional plan without being able to confirm the final arrangement. She might suggest selecting drawings in two matching formats for now. This decision does not require pretending that everything is already known. It requires explaining why the step is appropriate now and what it leaves open. Participants can disagree: restricting the format may change the very purpose of the exhibition for them. Incomplete information does not give a coordinator a special right to replace others' preferences with her convenience. A provisional answer still needs an intelligible scope.

The justification for a judgment cannot be assessed solely by the speaker's confidence. Lida might speak hesitantly while relying on a checked sample; another person might sound decisive while repeating an unconfirmed possibility. Expressing doubt therefore need not mean having weak grounds. It can describe their limits more accurately. However, a cautious tone proves nothing either. If someone repeatedly adds perhaps to a promise that others understand as accepted, an occasional verbal qualification does not clarify the agreement. What matters is the content of the answer and what follows from it.

Several kinds of confidence can coexist without an invented numerical scale. A measurement confirms the wall's size; access to the adjacent part of the room is under discussion; the final number of drawings depends on participants' choices. The information is uneven and can remain so. There is no need to assign convenient percentages to make the account look precise. For this shared work, it is more useful to distinguish observation, conditional agreement, and an answer not yet received. A number would need its own intelligible basis. Our imagined scene supplies no such basis.

IV. Admitting a gap and leaving people without guidance are different actions

After hearing that some dimensions remain unknown, participants still need to understand how to prepare their drawings. Lida could stop at identifying the gap and regard her contribution as complete. But if she agreed to coordinate preparation, people can expect at least an intelligible next inquiry. This does not impose a duty to remove uncertainty immediately. She can say when she will collect dimensions, who is willing to provide them, and which action remains unconfirmed. Continuing to take responsibility concerns organizing work on the question rather than producing reassurance at any cost.

The next action need not be a search for information. Sometimes the group needs to settle its purpose: will it show every selected work, or retain participants' option to decline a crowded arrangement? Sometimes changing the sequence is enough: obtain the selection before confirming placement. Sometimes it is more useful to acknowledge that the project still lacks a final agreement. Each answer can be honest if it describes the real situation. A promise to clarify becomes empty when nobody knows what is to be clarified or where an answer could come from.

Lida cannot command other people's time to fulfill her role. She can ask them to provide dimensions by an agreed date, but she cannot interpret silence as confirmation of the required format. No response still means no response. Shared preparation may nevertheless have deadlines of its own. The group then needs to discuss what happens to a drawing whose dimensions have not arrived: is its inclusion postponed, is the display changed, or is space retained? The condition should be intelligible in advance rather than appear as punishment for an assumed reason behind someone's silence.

A responsible next step can be small. Lida might collect one list of confirmed dimensions and mark two open entries. This is not a report proving her competence or a compulsory system for every meeting. In the imagined project, the list enables others to see which conclusion is available now. If no one needs that clarification, she can omit it. A record has value through the confusion it resolves, not through the number of documents produced. Working with uncertainty can itself become extra busywork that needs proportion.

V. A conditional answer must retain a real condition

Lida proposes holding the exhibition if the agreed drawings fit with the chosen spacing. That can be an intelligible conditional answer. Its meaning, however, depends on the sequence of actions. If participants already promise a complete display and give up another venue, the condition ceases to preserve much room for reconsideration. The words if they fit remain, while the consequences become final before the check. An honest conditional answer therefore requires examining what people will do before confirmation and which options they will genuinely retain afterward. The word if alone does not make a commitment reversible.

Sometimes the condition lies outside the coordinator's influence. Lida can confirm that she will submit a request without promising the room owner's agreement. She can report an agreement received without guaranteeing that circumstances will never change. These limits are not a reason to abandon every promise. She can answer for an available action of her own while describing precisely what depends on others. Uncertainty about someone else's answer should not be hidden inside her promise as though she simply preferred not to check her own part of the work.

A condition can also be unclear. If everything is fine allows everyone to supply a different understanding of fine. Shared work may benefit from identifying one relevant circumstance: access during stated hours is confirmed, the final list is agreed, or a particular display method is available. This does not require listing every possible obstacle. Such a list would be unattainable and could create another false confidence. It is enough to make intelligible the condition on which this particular answer actually depends.

Participants may prefer an arrangement independent of the missing information. They might choose to view drawings on a table rather than hang them all at once. That changes the purpose; it does not prove that the original plan was poor. Lida should name the change so that a convenient way around uncertainty does not appear as the old promise under a new label. If hanging the works mattered to someone's participation, that person can reconsider their agreement. Preserving responsibility includes recognizing the transition rather than declaring every available substitute sufficient for everyone.

VI. A check can be delegated; responsibility for the message remains

Lida asks an acquaintance to check the surface dimensions. He can measure them more accurately and is willing to help. Delegating the check is reasonable if both understand the question, conditions, and expected answer. But doing so does not immediately give the coordinator a checked result. She must wait for the response and preserve its scope: he measured an accessible wall at a particular time, not every possible use of the room. Another person's help adds support only within the action actually performed. Assigning a task is not evidence of its completion.

An unclear part of the response need not make a follow-up question an expression of distrust. The acquaintance says there is enough width, but Lida does not know how many sheets his calculation included. She can ask about the grounds for that conclusion and explain why they matter. She need not demand a full account of every movement of the measuring tape. A modest task calls for enough precision to relate the answer to the question. The degree of checking depends on concrete consequences and conditions. Our scene supplies no universal standard of supervision.

When passing the result to participants, Lida distinguishes her own observation from testimony received. She may identify who checked the dimensions and what he reported, where that attribution is appropriate and agreed. This does not mean looking for someone to blame in advance. The origin of an answer helps people know where to direct a new question and what knowledge is currently available. If the acquaintance has not agreed to share a personal detail, she should not add it to sound convincing. Usually the kind of check can be named without publicly exposing the helper or his private correspondence.

Responsibility for relaying information is limited too. Lida cannot guarantee that someone else's measurement is flawless, but she can preserve its meaning, disclose a known qualification, and correct her message if a discrepancy appears. If she stated a broader conclusion than she received, that concerns her own action. The fact that someone else performed the check does not fully explain it. Distinguishing contributions makes it possible to discuss a specific correction without blaming everyone who helped. It preserves cooperation in which nobody needs to pretend complete self-sufficiency.

VII. A new answer must be related to what has already been said

A few days before the exhibition, the adjacent wall turns out to be unavailable. Lida previously said that its use was still being discussed. She can now update the condition while preserving its former status: the possibility was not confirmed. But if she said the space would definitely be available, she needs to acknowledge a more specific discrepancy. Circumstances changed and the initial answer was unsupported are different explanations. Both may apply in one case, but one should not conceal the other. The review starts with what was actually known and communicated at the time.

Participants need to understand which actions the correction affects. Someone prepared an extra drawing expecting room; someone else agreed to help only under a particular display arrangement. Communicating the change gives them an opportunity to reconsider participation. It is not a promise to remove every inconvenience. Lida can propose an alternative and identify its limits without assuming that everyone must accept it to preserve the project. Correcting an answer matters beyond the speaker's inner honesty. It restores others' access to the condition on which they make their own decisions.

Sometimes a person wants to replace a correction with a detailed account of why they could not have known. That explanation may be true and still evade the main question: what now changes in the shared work? A brief reason can help clarify the answer's limit, followed by an available action. Responsibility is not measured by the volume of self-justification or intensity of self-blame. In this scene, promptly reporting that space is unconfirmed is more useful than spending hours deciding whether Lida is a generally reliable person. The present correction has a specific object.

Not every new fact proves a past mistake. After the exhibition, a participant might say he would have preferred a different arrangement. That is new information about his preference. If he expressed it earlier, the group should examine how it was considered. If not, later knowledge should not automatically be treated as previously available. Otherwise, every disappointment becomes an accusation that the responsible person ought to have anticipated someone else's experience. Learning from the comment remains possible without that demand. A future agreement can be refined without rewriting the history of earlier information.

VIII. Pressure to be certain does not create grounds

In our story, one participant asks Lida simply to confirm everything now: people are tired of discussing it. Their fatigue is real, but it does not mean the open question is resolved. Lida can propose a narrower option, agree on the timing of another answer, or decline to give final confirmation. Others may find that refusal unpleasant. It leaves uncertainty they wanted to end. Yet an unsupported confident message only transfers that uncertainty into the hidden part of other people's actions, where it is harder to notice and discuss.

The coordinator may also hold power. If Lida alone relays information, she can use uncertainty to retain control of decisions—for example, indefinitely saying that conditions are being clarified without sharing an answer already received. Unknown information does not justify that treatment of known information. Responsibility requires distinguishing the actual remaining gap from what can already be passed on. A delay alone does not authorize us to attribute manipulation. We identify an action that can be discussed: which confirmed information was withheld and why it matters to participation.

Sometimes the conversation itself becomes unsafe. A person is required to prove competence by disclosing private matters, threatened with consequences for expressing doubt, or prevented from declining an expanded role. These distinctions do not impose a duty to continue that discussion. Someone may stop the exercise, limit a message to what is necessary, or leave voluntary participation where that is possible. A threat does not become a sound condition for examining knowledge merely because the question formally concerns the project. Safety and personal boundaries may matter more than a complete explanation under pressure.

The right to acknowledge uncertainty does not oblige others to stay in a project without the answer they need. They may consider the uncertainty unacceptable and decline participation. In a small voluntary undertaking, that can honestly end an agreement. A responsible person acquires no right to demand trust in place of confirmation. Her task is to make conditions intelligible within the accepted role. Someone else's resulting choice may differ from what she wanted. Naming the knowledge boundary supports participants' autonomy only if they are genuinely free to use the information in their own way.

IX. Inquiry needs a scope, not a promise to learn everything

Lida could continue checking almost indefinitely: learn every visitor's habits, predict the convenience of every arrangement, and discover in advance who will change their mind about participating. Those questions differ in accessibility and relevance to the present step. Agreeing on placement requires one set of information; guaranteeing that everyone will enjoy the exhibition would require unattainable knowledge. Admitting that limit does not make preparation pointless. It makes it possible to stop promising the kind of answer the chosen check cannot provide in principle, even when performed carefully.

A present question may be settled at one level while future experience remains open. Lida confirms an agreed number of places for known formats. She does not claim that every drawing will be equally well received or that participants will retain their current view of the exhibition. This limit is not small print added after a promise. It is part of the answer itself. If others require a different degree of confirmation, that needs discussion before action rather than treating a broader meaning as naturally included in every positive message from the coordinator.

Timing needs distinction too. Further checking may be useful, yet its answer may arrive too late for an available decision. Alternatively, the deadline can be moved and the rush merely reflects a familiar schedule. Lida can identify both what remains learnable and what changes while she waits. This is not a formula for the optimal value of information. It does not determine acceptable risk for someone else's participation. The discussion simply restores time to the agreement so that later does not conceal the consequences of waiting already taking effect.

If a significant condition cannot be checked through an available method, lack of confirmation can be acknowledged as the final result of the present inquiry. That is not the same as a negative answer about the condition itself. It means the project cannot currently rely on that condition as established. The next possibilities include changing the purpose, waiting on an agreed basis, or ending participation. Each has its own cost. A coordinator can help identify those costs without turning the remaining gap into proof that everyone must choose her preferred way out.

X. An adequate answer can be more honest than an impeccable image

Lida finishes preparation without finally knowing everything about the exhibition. She knows which drawings are agreed, which placement has been checked, who confirmed participation, and which experience still belongs to the future. In this scene, that is enough for a small voluntary project. Another task could have stricter requirements and a differently defined role. The story therefore yields no universal recommendation to check less. It offers a narrower question: does the message correspond to the grounds and consequences to which it actually applies?

The image of a competent person can blend several qualities: possessing information, answering quickly, staying calm, and never correcting a statement. They need not coincide. A correction can require more responsibility than preserving a confident account. A pause can help clarify a question or conceal inaction. It is more useful to discuss the content of an answer and its next step than to infer someone's quality from conversational style. Participants need intelligible conditions of cooperation, not an impeccable public role for whose sake a significant gap stays invisible.

Acknowledging uncertainty need not become a permanent way of presenting yourself. You can know your part of the work well while answering one question with insufficient information. You can decide on a stated assumption and later reconsider in light of a new message. You can decline an expanded commitment that was never agreed. These are different actions with different grounds; a general cautious phrase does not replace them. What matters is maintaining correspondence between what is known, what is said, and what is promised, rather than appearing invariably modest or decisive.

After the exhibition, Lida may see that clarifying one question helped people agree while another remained unclear. That is material for the next undertaking, not an assessment of her whole personality. She can retain a useful way of communicating, shorten an unnecessary record, and decline a duty to predict everything. A responsible role remains human: it includes personal capabilities, others' contributions, and changing circumstances. Admitted uncertainty does not automatically complete the work. It gives the work a precise object when the person distinguishes available information, the scope of commitment, and an action possible after the answer.

What this means in practice

If it suits you, choose one small, safe question on which someone expects an answer from you in voluntary shared work. An imagined scene is enough; you need not accept a real commitment for the exercise. Avoid questions involving treatment, law, financial transactions, or other serious consequences. You can decline without explanation, keep private matters private, and stop if there is a threat, coercion, or unwanted pressure.

Identify the question and your agreed role. Then distinguish what is confirmed, what is assumed, and which answer is missing. Formulate one limited response that does not imply a broader promise. If the question remains your responsibility, identify an available next action or acknowledge that you currently lack a way to clarify it. Do not assign a check to someone without consent or treat handing over a question as obtaining a result.

Consider what action another person might associate with your response. If it requires missing confirmation, you can narrow the answer, change the sequence, or leave the decision open. You need not send your formulation or secure anyone's agreement. The exercise's modest result is a clearer boundary between knowledge and commitment. It is not a test of honesty, a guarantee of trust, or proof that any decision under uncertainty is now safe. Where circumstances are significant or exceed your role, the record cannot replace necessary specialist information and actual agreements.

Key quotation

demonstrated or demonstrable

Thomas Henry Huxley, “Agnosticism” (1889), Volume V, p. 246; a brief fragment. The author warns against presenting unproved conclusions as certain. These words concern the grounds of a claim, rather than a promise to prove any everyday answer.

Connections between traditions

This comparison is limited to historical philosophical distinctions and the authors' contemporary reasoning about cooperation. Locke examines the grounds for judgment without certainty; Huxley distinguishes a method of examining grounds from claiming certainty. Our scene asks how to communicate an answer's limits to people relying on it. These are different questions. Philosophical argument is not an experiment on trust, and a voluntary exercise gains no demonstrated effectiveness from a classical reference. Huxley's religious polemic remains the source's context; its historical judgments about faith and churches are not adopted as this article's conclusions or conditions of reader participation.

Sources

Thomas Henry Huxley. Agnosticism (1889), Collected Essays, Volume V, Science and Christian Tradition, pp. 209–262; Huxley File digital primary transcription. Actual reading covered the opening and selected continuous context concerning method and grounds, including pp. 246–247. Neither the complete essay nor volume was read. Historical philosophical polemic, not empirical research on our adult scenes. No open licence for the digital edition was established; worldwide territorial rights status is not asserted.
John Locke. An Essay Concerning Human Understanding, Volume 2. Project Gutenberg 10616, based on the second edition, Books III–IV; digital version updated October 28, 2024. Actual reading covered metadata and contents, all of Book IV Chapters XIV–XV, and the selected opening of XVI, sections1–4, in displayed contexts. The whole Essay was not read. A narrow philosophical reference, not research evaluating this exercise. Gutenberg's license includes territorial conditions; worldwide public-domain status is not asserted here.

About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, unity across traditions, intercultural dialogue, and contemporary spiritual discipline. Instagram: @andrey.zarinsh.

Maria Evilkut is a co-author of United Wisdom. Alongside her research, she seeks to test these ideas in practice and observe how principles of unity and mutual understanding work in real life between people and cultures. Instagram: @ilkutki.