What Counts as Evidence: Observation, Records, and Experiment

Epigraph

There remains but mere experience, which, when it offers itself, is called chance; when it is sought after, experiment.

Francis Bacon, Novum Organum I.82. Exact historical English sentence in the edition edited by Joseph Devey; the following context calls for ordered inquiry.

What this article is about

Someone saw a closed door. A message says a meeting was cancelled. After a timetable change, several participants arrived on time. Each observation may matter, but each answers a different question. The door shows its condition at a particular moment; the message preserves what was said; changing conditions may test an explanation if the check is appropriately designed. No kind of material is automatically stronger than every other. This article considers how observation, records, another person’s testimony, and experiment become grounds for a specific claim. Historical passages from Bacon, Hume, and Dewey offer different philosophical starting points rather than replacing modern research standards. The examples are invented, and the exercise examines one everyday report while respecting privacy and refusal. We ask both what was found and what it can establish. Confirmation, correction, and an honestly unresolved question are all possible outcomes.

I. Evidence is evidence for a question

Begin with “The meeting was cancelled.” Are we establishing the organizer’s decision, notification of participants, or the absence of people from a room? These circumstances connect without being identical. A decision may be made but not announced; a notice may be mistaken; a closed door may mean the group moved elsewhere. Name the question before collecting material. Otherwise discussion gathers everything that looks persuasive while leaving unclear what has actually been established. A precise claim need not be lengthy. Identifying the event, time, and relevant conditions lets others understand what the available information is supposed to support, and where support might still be missing.

At the opening of How We Think, Dewey distinguishes beliefs whose grounds have not been considered from beliefs whose grounds have been examined [1]. This is a philosophical and educational account, rather than a modern validated method for every field. Its useful question concerns the connection: material supports a conclusion when we explain why it supports that particular conclusion. Seeing a message and concluding that everybody received it are different steps. The second needs additional grounds. Attention moves from a story’s overall persuasiveness to the passage from available material to the claim being made.

A photograph can support several different claims. If authentic and correctly associated with a location, it shows that a bench was in the frame. Establishing when it was taken requires another check; why people were absent is usually not visible by itself. These limits do not make the image useless. They give it a defined place in the argument. “This image shows an empty room; its capture time has not been established” preserves the available information without turning the picture into a witness to every desired circumstance. The distinction concerns the question, rather than a moral judgment of the person who supplied the photograph.

Evidence does not have to end disagreement immediately. It may separate two questions, locate the disagreement, or make one explanation less suitable. If an organizer confirms writing a cancellation notice, that helps establish the message; whether the decision later changed remains another question. Material can advance understanding without completing it. Distinguish “relevant,” “supports this part,” and “sufficient for this conclusion.” These phrases are not a mandatory universal scale. They describe the specific contribution of information without demanding either absolute reliability from every fragment or a final answer to the entire situation. Partial progress can be stated precisely rather than disguised as certainty.

II. Observation has a position and a scope

Observation gives contact with a circumstance, but that contact has a time, position, and method. Someone saw a door from the street at ten, rather than everything inside the building throughout the day. Stating the scope does not belittle an eyewitness; it makes their report intelligible. Ask what was visible, what was directly heard, and what was subsequently inferred. Courteous precision leaves “I did not notice” available. An expected detail should not be supplied merely because it completes the explanation neatly. Not noticing a detail and observing that a detail was absent may require different conditions, and the report should preserve whichever actually occurred.

In Section X of An Enquiry Concerning Human Understanding, Hume considers how trust in testimony depends on experience and the circumstances of a report [2]. We use that bounded question without importing his entire argument about miracles. Historical remarks about delivery should not become an established way of detecting deception: hesitant speech can have many causes. Our analysis asks instead about access to the event, the report’s content, and its relation to other material. A specific statement is examined. Vocal confidence alone neither confirms nor refutes what the person says happened.

Compare two reports of a closed door. One participant passed by; another checked the entrance and saw a notice. The second knows more about the entrance, but not necessarily more about the decision’s reasons. The first may have noticed people gathering at a neighbouring door. One difference in access should not give people permanent titles of “reliable” and “unreliable.” Separate each report into parts and identify their support. When accounts conflict, clarify the time and subject before demanding an admission of error. Different events or stages do not justify every account, but they deserve examination when the known circumstances leave that possibility open.

Observation may be repeated when appropriate and permissible. Checking whether a door is open now is easier than establishing its condition yesterday. A new observation does not rewrite the earlier event; it answers a present question. The past requires traces, records, or testimony with suitable origins. Everyday clarification does not justify surveillance or concealed recording. Another person’s privacy remains relevant even when stronger evidence would be convenient. Without permitted access, retain a limited report or decline the conclusion. An adequate approach considers both what can be seen and whether the way of obtaining that information respects the people involved and the setting’s actual boundaries.

III. A record preserves a report and needs provenance

Records are useful because people can revisit them and show them to someone else. But a “record” may be an original message, a copy, a later extract, a corrected table, or a summary. Their connections to the event may differ. A screenshot shows a displayed fragment while sometimes hiding its date, sender, or neighbouring messages. Full context need not be publicly exposed; the relevant portion can be checked through permitted access. Do not attribute an original’s properties to an extract without grounds. A preserved text becomes more useful when its creator, time, purpose, and subsequent changes are sufficiently clear for the question at hand.

Imagine two timetable versions, one starting at ten and another at eleven. A file’s save date does not necessarily date the decision: an old text may have been copied later. Establish the version’s connection to the agreement rather than selecting it by appearance. An organizer’s change notice may help; participants’ acknowledgments answer the different question of who learned about it. A small clarification need not reconstruct the entire correspondence history. Identify the event and find material that can distinguish the versions substantively. If none is available, “the last file found” describes the search, rather than establishing that it contains the most recent decision.

Several records may look like independent confirmation while repeating one source. Three forwarded messages can descend from a single mistaken entry. Their number shows circulation, rather than three separate checks of the event. Repetition is not always useless: it may establish who could access a text or how the wording changed. Distinguish a report’s circulation from the independence of its grounds. State a common origin plainly, then seek another relevant route if the conclusion needs further support. One accurately described piece of evidence can be more intelligible than numerous links presented as proof without explaining the chain they share or the separate question that chain can answer.

Records can preserve intentions, rules, and requirements as well as events. An established rule text shows the wording adopted or published; compliance requires other information. A meeting plan does not prove the meeting took place. A completion report is useful, and its content can be compared with the result when the question requires it. This is not blanket distrust of writing. It respects writing’s particular function. A recorded promise to open a room supports the claim that the promise was made; the door’s condition concerns something else. Separating these levels makes responsibility discussable without treating every document as either an infallible seal or an inherently empty piece of paper.

IV. Experiment asks through changed conditions

Everyday conversation often calls any new action an experiment. A meeting time changes, the outcome is pleasing, and someone declares the cause discovered. Change alone does not establish causation. Specify the explanation being tested, what changes, what is observed, and which other circumstances might account for the result. A small trial can be useful without being a rigorous study. That usefulness survives an honest description of its scale. “This time suited the participants on this occasion” differs from “This time always improves attendance.” The first reports a particular outcome; the second extends beyond one meeting and needs a different evidential basis to support that extension.

In Aphorism I.82, Bacon distinguishes experience encountered by chance from experience deliberately sought, then calls for an ordered course of inquiry [3]. This is a historical programme for investigating nature, rather than a complete modern experimental standard. Its question matters here: why arrange the trial, and what should the result clarify? If time, venue, invitation, and participants all change, a pleasing outcome does not isolate one change as the cause. The new combination can be acknowledged as workable while the narrower explanation stays open. This preserves the outcome without granting it more evidential force than the comparison supplies.

Suppose a group voluntarily compares two reminder methods. Agree beforehand on what participants are willing to do, what information may be recorded, and how results will be described. Silence should not automatically count as consent. Do not conceal the actual purpose or create inconvenience to obtain a more striking conclusion. If conditions are too different for comparison, record that limitation rather than artificially equalizing people. The trial remains a bounded shared arrangement, rather than grounds for a universal recommendation. A participant may decline or suggest another contact method. Curiosity does not authorize making another person’s participation into material collected regardless of their willingness or the agreed purpose.

Experiment is sometimes impossible or inappropriate. A past event cannot be recreated as the same event, and an unsafe possibility should not be tested on people merely for information. Inquiry then uses available observations, records, comparisons, and testimony with stated limits. An unavailable experiment neither permits every conclusion nor cancels all other knowledge. Choose an approach suited to the subject and permissible in the situation. In a specialist field, an everyday trial cannot replace that field’s qualifications and requirements. The present discussion concerns philosophical clarity about grounds, rather than instructions for research that could put someone’s health, property, or access at risk.

V. Different materials need explained connections

There is no need to choose permanently between observation and records. A record may establish an agreement, observation the room’s condition, and a participant’s account an unrecorded circumstance. Their combination helps when the parts address specific questions. Different kinds of material cannot simply be added together and the total called proof. Explain the connections: how the time is known, why this entry concerns that meeting, and what connects an empty room with cancellation. If one connection is unestablished, other information may remain correct. Preserve its value while identifying where the conclusion has extended beyond the support actually available for that particular step.

Suppose a cancellation notice was sent at nine, the door was closed at ten, and someone reports a move to a neighbouring room. Clarify when the move happened, who was notified, and whether the original cancellation still applied. There is no need first to decide who is “generally right.” One person may accurately describe an early decision and another a later situation. An established sequence can resolve the substantive conflict; an unestablished sequence remains only a possible explanation. A well-formed story does not supply evidence for its connections. “Perhaps this happened” should remain distinct from an established sequence, even when the suggestion is convenient to everyone involved.

Notice where information in a chain was obtained directly and where it was relayed. “I saw a notice” differs from “Someone told me there was a notice.” A relay can be conscientious and important, but it adds a link. If the original is unavailable, do not imply access to it. Attribute the account and state the scope of your check. This matters when information travels further: the next person should understand whether they receive an observation, an extract, or a summary. Clarity does not require publishing everybody’s personal details. Describe origin to the extent needed for evaluating the conclusion and permitted by the relevant privacy arrangement.

Combining material should not create an appearance of omniscience. Even a well-established meeting time does not explain every intention or experience. If the conversation turns to fairness, convenience, or personal meaning, identify the new question. Factual clarity helps discuss it on appropriate grounds without selecting values for people. A late notice can be established and an acceptable notification arrangement discussed separately. Learning that a time did not suit someone does not establish a hidden intention to disrupt the meeting. Distinguishing subjects connects knowledge with decisions while preventing one established detail from becoming an explanation of the entire human situation surrounding it.

VI. Absence and conflict need an appropriate check

“I did not find the message” reports the outcome of a particular search. Concluding that no message existed requires knowing where it should have been and how completely that place was checked. Searching one chat does not cover another channel or a personal conversation. A limited check may suffice for the next action, such as requesting the notice again. Preserve its limits when explaining the past. This does not demand endless searching. Stop with “It was not found in the accessible place checked” and make a decision that acknowledges the uncertainty. Failure to find evidence and evidence of absence differ through search conditions, rather than the forcefulness of the conclusion.

A conflict first needs clear formulation. “The door was closed at ten” and “The meeting happened” need not contradict each other. “The group occupied this room at ten” and “This room was empty at ten” look more directly incompatible if location and time have been established consistently. Clarification should not become endless rescue of a preferred account. When conditions genuinely coincide, acknowledge the conflict and identify what can be checked again. The cause might be an incorrect time, a mistaken room, or continuing disagreement. Bad faith need not be declared immediately, but incompatible reports should not be concealed beneath general language about people having different perspectives.

An unsuccessful check may reveal something about the check rather than only about the explanation. A cropped record lacks the required date; an observer who arrived late did not see the beginning. Such limits do not automatically confirm the opposite account. “This material does not establish cancellation” differs from “Cancellation definitely did not occur.” State the open question and what could clarify it. If another attempt would not improve access to the relevant circumstance, repeating the same procedure has little purpose. Limits can be recorded without accumulating identical actions. A sound stopping point explains its grounds and leaves revision available when genuinely new material appears.

Keep a negative result when it concerns the question. Participants checked the announced venue and found no group; that information matters even if a relocation later emerges. A report of a successful new arrangement should retain occasions when it did not suit people. This makes the description more specific and its conditions visible. One difficulty does not necessarily reject the whole arrangement. Compare it with the actual claim: “suits some meetings” and “always suits everybody” have different scope. The claim may be narrowed, the arrangement changed, or the proposal abandoned. That choice should follow substantive analysis rather than the desire to preserve an attractive story about success.

VII. Enough information and permissible action differ

Not every everyday task needs a final explanation of the past. Arranging the next meeting may require only clarifying its time now. Correcting a mistaken report requires identifying which part was wrong. The decision depends on the question and consequences. Yet the need to act does not automatically make insufficient information sufficient. A cautious, reversible step can be chosen while acknowledging that yesterday’s failure remains unexplained. This preserves both practical possibility and accuracy. “We do not yet know, but we can clarify the next arrangement” is a substantive outcome when it matches the available circumstances and does not conceal a material risk to participants.

A consequential action needs appropriate grounds and competence. A small timetable check should not become a model for decisions about health, law, or safety, where specialist requirements apply and errors have a different scale. Even an ordinary group should consider who bears a decision’s consequences. An organizer’s convenience is not everybody’s agreement. A fact can be confidently established yet used unfairly. “Do we know enough?” therefore connects with “Are we entitled to act this way?” and “Who participates in choosing?” These questions belong together, but neither a record nor an observation supplies everybody’s final answer about an action’s permissibility or the distribution of its costs.

Sometimes people disagree about how much uncertainty they can accept rather than about the information itself. One will try a new time for a single meeting; another cannot risk travelling without clear confirmation. The second need not be judged less open to experience. Their participation conditions differ. Offer separate confirmation, another contact method, or the option not to attend. A shared decision becomes more concrete when factual grounds and uncertainty’s practical cost are distinguished. Easy agreement is not promised: constraints may remain incompatible. They can still be named without diagnosing someone’s character or requiring participation in an inconvenient trial as proof of good faith.

Checking should not become an endless examination of another person. If an ordinary clarification continually demands more private information, reconsider the conversation’s arrangement. Define the relevant question, the minimum material needed, and a stopping point. A person may decline to disclose personal context unrelated to the shared task. Refusal does not automatically make their report false. Another participant may also decline to rely on information they cannot adequately check. Limits of knowledge and limits of access can coexist. A decision can acknowledge both instead of applying pressure for more disclosure or demanding an unconditional promise to trust whatever has been reported.

VIII. Reporting a result preserves its scale

After checking, report what was done and what was established. “I checked two accessible messages; the later one gives a new time” is more precise than “Everything is verified.” A limited check is not negligible: its value lies in the question it resolves. State separately what remains unknown, such as who received the notice, whether the meeting happened, or why it moved. The next reader can then assess both the conclusion and its boundaries. Brevity does not require hiding conditions. A single sentence may prevent a found record from becoming complete knowledge of an event, or a particular result from becoming guaranteed reliability of an entire arrangement.

When a correction appears, preserve the distinction between the earlier report and the new one. Do not silently replace a date as though no mistake existed when others already relied on it. Name the correction briefly and inform those affected. Accuracy does not require public humiliation of the person responsible. “The earlier notice said ten; the confirmed time is eleven” explains the change without judging a personality. If circumstances changed and the earlier record was correct at the time, say that too. Correcting an error and announcing a new arrangement need different wording, even when both produce the same next meeting time.

Leave room for a new check without permanent doubt about everything. A time confirmed by an appropriate message can be used until a relevant change appears; every detail need not be established anew each day. New information is examined substantively: does it concern the same event, alter a condition, or require a correction? This avoids both careless fixation and endless accumulation of procedures. Knowledge gains a workable stability connected to its subject and available grounds. This is an original proposal for everyday handling of information, rather than a promise to eliminate error. People retain the ability to act and responsibility to clarify a result when a significant discrepancy emerges.

Observation, records, and experiment need not compete for the sole title of genuine evidence. They offer different approaches to a question, whose possibilities and limits can be combined when connections are explained. A material’s prestige matters less than its relation to a particular conclusion. That relation may be strong, partial, or still unestablished. A clear report preserves all three possibilities. It allows respect for another person’s testimony alongside checking, and uncertainty alongside action. The intended outcome is modest: a reader understands what is claimed, what supports it, what the support does not establish, and which next step remains appropriate and voluntary for the people concerned.

What this means in practice

Choose one everyday claim that matters for a next action and does not require another person’s private information. State the subject, time, and conditions: “The accessible notice schedules the meeting for Wednesday.” Do not yet add that everybody was notified or that the meeting already happened. Separate those questions if they also matter.

Name the material available: direct observation, an original record, a copy, a relay, or the outcome of an agreed trial. State its origin and what was checked. A record needs relevant provenance and version; an observation needs time and access. If several reports repeat one source, do not count them as separate checks.

Explain the connection between material and claim in one sentence. Ask which part exceeds its support. A record might establish the scheduled time without establishing that everybody received it. If accounts conflict, first compare location, time, and wording. If no material was found, describe the search’s scope rather than claiming absolute absence.

Choose one permissible next step: ask a participant, check an accessible original, correct the report, or leave the question open. Do not arrange an experiment involving others without agreement. Set a stopping point proportionate to the question. If another search cannot improve relevant access, stop and state the limitation.

Report the outcome within the check’s scope: confirmed, corrected, or unestablished, with brief grounds. State a decision separately if one was made. This original exercise is neither a demonstrated psychological intervention nor a universal research protocol. A sufficient outcome is a more precise report and an intelligible next step, rather than guaranteed agreement or final truth about the entire event.

Key quotation

others are accepted because their grounds have been examined.

John Dewey, How We Think I.1. Exact contiguous clause from a sentence contrasting beliefs whose grounds have or have not been considered; lowercase start retained.

Connections between traditions

Question	Primary grounding	Limit of comparison
Were a belief’s grounds examined?	Dewey, I.1 [1]	A method of obtaining belief differs from the result’s truth.
What is observed and what inferred?	Dewey, I.2 [1]	A sign’s connection with a conclusion needs explanation.
How should a relayed report be assessed?	Hume, X.88–89 [2]	Historical reasoning does not supply a deception test.
Why change conditions?	Bacon, I.82 [3]	Deliberate inquiry still needs order; an everyday trial stays bounded.
How should a found record be handled?	Editorial application of questions [1–3]	Provenance, version, and execution are separate; this is not a quotation from the classics.

This is an editorial comparison. The three philosophical positions do not merge into a modern protocol or demonstrate the proposed exercise’s efficacy.

Primary sources

John Dewey, How We Think. 1910. English original, Project Gutenberg #37423. Text.
David Hume, An Enquiry Concerning Human Understanding. 1748; selected reprint1902. English original, reprinted from1777 edition, edited L.A.Selby-Bigge, second edition1902, Gutenberg #9662. Text.
Francis Bacon, Novum Organum. 1620; selected English edition1902. Historical English version edited Joseph Devey, P.F.Collier & Son1902, Gutenberg #45988; translator not identified by consulted title page. Text.

Actual reading is bounded: Dewey I.1–2 selected grounds and observation/inference passages; Hume X Part I paragraphs 88–89 with the opening of 90; Bacon I.19,46,49 and selected82 on experience and ordered inquiry. These are philosophical primary texts, not contemporary experimental studies. Dewey’s Columbus/flat-earth illustration is not adopted. Bacon’s English is the 1902 edition edited by Joseph Devey; the consulted title page does not identify a translator. Short Russian quotations are United Wisdom’s own translations of the specified English text. Examples, record-provenance applications, and practice are original; neither whole-book reading nor demonstrated efficacy is claimed.

About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, the unity of 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.