What Counts as Evidence?

INTRODUCTION: EVIDENCE IS NOT THE SAME AS PROOF

Introduction: Evidence is not the same as proof

People frequently say:

• “There is evidence for it.”
• “There is no evidence.”
• “This artifact proves it.”
• “The ancient text says it happened.”
• “Scientists agree.”
• “Historians are hiding the truth.”
• “The absence of records proves a cover-up.”

Each statement can sound decisive while leaving the important questions unanswered.

Evidence for what exact claim?

How directly does it bear on that claim?

Who produced it, when, and for what purpose?

Is it independent of the other evidence being cited?

What alternative explanations fit the same facts?

How much of the conclusion comes from observation, and how much comes from interpretation?

A clay tablet is evidence. A museum record is evidence. A laboratory measurement is evidence. A royal inscription is evidence. A modern forgery is also evidence—but perhaps evidence for modern collecting, fraud, ideology, or the antiquities market rather than for the ancient history it pretends to record.

Evidence does not speak without questions.

The Archive therefore begins Course 2 with a rule that governs every later course:

Evidence is information that changes how reasonable a claim is. It does not determine the conclusion automatically.

01 · COURSE METHOD

The six parts of a historical argument

A responsible historical argument separates six elements.

Claim

A claim is a proposition that can be supported, weakened, or contradicted.

Examples:

• This tablet was written in the seventh century BCE.
• This composition existed before Genesis reached its final form.
• This king fought the campaign described in his inscription.
• This object came from Nineveh.
• This image depicts a spacecraft.
• This flood story preserves memory of a historical disaster.

A useful claim must be precise enough to test. “This is mysterious” is not yet a historical claim.

Source

A source is the object, text, image, dataset, testimony, record, or observation from which information is drawn.

Examples include a clay tablet, excavation diary, museum accession record, ancient inscription, modern translation, photograph, satellite image, laboratory report, scholarly article, or oral history.

Observation

An observation states what can be directly detected or documented.

Examples:

• The tablet is made of clay.
• Three lines are broken.
• The museum assigns the number K.3375.
• The inscription names a ruler.
• The symbol has four visible projections.
• The object entered the collection in 1882.

Observations should be described before meaning is assigned.

Inference

An inference connects observations to a proposed explanation.

Examples:

• The handwriting resembles Neo-Assyrian forms.
• The damage probably occurred before excavation.
• The text was copied from an earlier exemplar.
• The symbol may represent a star.
• The object may come from an undocumented excavation.

An inference can be strong or weak. It should not be disguised as a direct observation.

Interpretation

An interpretation explains significance within a larger historical, literary, political, religious, or cultural framework.

Examples:

• The inscription legitimizes royal authority.
• The flood account criticizes divine disorder.
• The image represents kingship rather than astronomy.
• The text transforms an inherited tradition.
• The object’s display history reflects colonial collecting.

Conclusion

A conclusion states what the full body of evidence justifies.

A good conclusion includes:

• the exact claim;
• the evidence supporting it;
• the limits of that evidence;
• competing explanations;
• and the appropriate level of confidence.

Archive standard

Never allow the conclusion to masquerade as the observation.

02 · COURSE METHOD

What is a primary source?

The Library of Congress describes primary sources as original documents and objects created during the period under study. The category also varies by discipline and research question.

A source is primary for a particular question.

A seventh-century BCE copy of Gilgamesh is:

• a primary source for Neo-Assyrian scribal practice;
• a primary source for the wording preserved in that manuscript;
• evidence that the composition existed by that time;
• but not an eyewitness record of a historical Gilgamesh living many centuries earlier.

A nineteenth-century excavation diary is:

• a primary source for the excavation;
• a primary source for the excavator’s methods and assumptions;
• but not necessarily a primary source for the ancient event being investigated.

A modern conspiracy video is:

• not a primary source for an alleged ancient event;
• but a primary source for modern belief, rhetoric, and reception history.

Primary does not mean true

A primary source may be:

• mistaken;
• propagandistic;
• damaged;
• fictional;
• symbolic;
• forged;
• copied centuries later;
• written from limited access;
• shaped by genre;
• or designed to persuade.

Royal inscriptions present victories from the ruler’s perspective. Legal texts may preserve ideals rather than uniform daily practice. Myths reveal ancient thought and literary tradition even when they are not modern documentary reports.

Evidence verdict

Established: Primary sources are original materials relevant to the period or phenomenon under study.

Methodological rule: Primary status does not guarantee accuracy, neutrality, completeness, or literal reporting.

03 · COURSE METHOD

Secondary and tertiary sources

Secondary source

A secondary source analyzes, interprets, compares, or explains primary evidence.

Examples:

• peer-reviewed article;
• critical edition;
• scholarly monograph;
• archaeological site report;
• historian’s synthesis;
• linguistic analysis.

Secondary sources are not inferior merely because they are secondary. A trained specialist may compare many objects, languages, contexts, and prior studies that no single ancient source contains.

The value of a secondary source depends on:

• relevant expertise;
• transparent method;
• access to evidence;
• engagement with prior scholarship;
• quality of reasoning;
• and correction by other specialists.

Tertiary source

A tertiary source summarizes existing knowledge.

Examples:

• encyclopedia;
• introductory textbook;
• museum overview;
• database summary;
• popular timeline;
• general reference page.

Tertiary sources are useful for orientation but rarely sufficient for a disputed claim.

Archive source hierarchy

For a major historical claim, prioritize:

1. the object or text itself;
2. excavation and collection records;
3. a current critical edition or official corpus;
4. specialist peer-reviewed scholarship;
5. institutional synthesis;
6. responsible public scholarship;
7. popular summaries;
8. unsourced repetition.

This is not an automatic verdict. A catalog can be outdated, and a public researcher can identify a genuine error. The hierarchy indicates how much checking is required.

04 · COURSE METHOD

The evidence chain: from object to reader

When a reader sees an ancient sentence in English, a long chain may stand behind it.

Stage 1: Ancient object

The physical object may be a tablet, papyrus, parchment, stone inscription, seal, pottery sherd, metal plaque, wall painting, bone, building, or another material remain.

Stage 2: Archaeological or collection record

Researchers document findspot, layer, associated objects, orientation, condition, dimensions, ownership, acquisition, and movement into a collection.

Stage 3: Image, drawing, or hand copy

The object may be represented by photograph, scan, reflectance imaging, three-dimensional model, line drawing, hand copy, squeeze, or tracing.

Each form preserves some information and loses other information.

Stage 4: Sign identification

A specialist identifies marks as signs, letters, words, damage, corrections, or later additions.

Stage 5: Transliteration

A transliteration represents ancient signs in a modern scholarly system. It is not yet a translation.

For cuneiform, one sign can permit more than one reading. Transliteration therefore already contains interpretation.

Stage 6: Transcription or normalization

A transcription may represent the presumed language rather than each visible sign. Terminology varies by discipline, so each edition’s conventions must be checked.

Stage 7: Translation

A translation renders meaning into another language.

Translation requires choices about vocabulary, grammar, idiom, damaged text, genre, word order, ambiguity, and cultural concepts.

Stage 8: Critical edition

A critical edition may combine multiple manuscripts, record variants, restore missing words, and explain editorial choices.

The smooth text a reader sees may never have existed on one complete ancient object.

Stage 9: Interpretation

A scholar or presenter explains what the text means and how it relates to other evidence.

Stage 10: Public retelling

A documentary, podcast, post, or website may condense the interpretation again.

Archive standard

The farther a claim moves from the physical object, the more important it becomes to preserve the chain connecting each stage.

05 · COURSE METHOD

Object, inscription, and claim are different questions

Three questions must remain separate.

Is the object ancient?

Material analysis, manufacturing technique, weathering, excavation history, and comparison may establish that an object is old.

Is the inscription ancient?

An ancient object can receive a modern inscription. Forgers can reuse ancient blank material. The object’s age does not automatically date the writing.

Is the statement historically accurate?

A genuine ancient inscription may exaggerate, omit, idealize, ritualize, or invent.

Authenticity establishes that ancient people made the statement. It does not establish that the statement is complete or factually reliable.

Suppose a stone says a king destroyed a city. Separate propositions include:

• The stone is ancient.
• The writing is ancient.
• The inscription was commissioned under the king.
• The king claimed to destroy the city.
• Archaeology shows destruction in a compatible period.
• An independent source attributes the event differently.
• The city was damaged but continued to exist.

Evidence verdict

Strongest direct conclusion from a genuine royal inscription: The ruler or institution wanted the claim publicly presented.

Further historical conclusion: Requires corroboration from archaeology, chronology, and independent records.

06 · COURSE METHOD

Archaeological context

An artifact from a controlled excavation is more than an isolated object.

Its context may include:

• coordinates;
• depth;
• stratigraphic layer;
• building phase;
• room;
• burial;
• associated ceramics;
• organic material;
• destruction layer;
• and relationship to nearby objects.

Context can help answer:

• When was the object deposited?
• Was it used in a household, temple, workshop, tomb, archive, or dump?
• Was it moved in antiquity?
• Was it found with datable material?
• Was it reused?
• Does it belong to the building where it was discovered?

An object without context may still reveal material, style, language, iconography, or technology. It loses much of its ability to establish original location, social function, association, chronology, and event.

Looting destroys knowledge

Illegal excavation does not merely remove an object. It destroys relationships among objects, layers, architecture, human remains, and environmental evidence.

Professional archaeological organizations and UNESCO emphasize that illicit trafficking and undocumented antiquities damage both cultural heritage and the information archaeological resources contain.

Archive standard

An object without context may retain market value while losing historical resolution.

07 · COURSE METHOD

Provenience and provenance

Provenience

In archaeology, provenience usually means the precise location where an object was found: trench, locus, layer, room, burial, or coordinates.

Provenance

Provenance generally means the history of ownership, custody, collection, sale, transfer, export, and acquisition.

A provenance record may include:

• excavator;
• dealer;
• collector;
• auction;
• export permit;
• museum accession;
• previous publication;
• and gaps in ownership.

Some disciplines use the terms differently, so definitions should be stated.

Why provenance matters

Provenance helps evaluate:

• authenticity;
• legal ownership;
• ethical acquisition;
• forgery risk;
• modern alteration;
• and whether a claimed findspot is documented.

An incomplete provenance does not automatically make an object false. It does reduce confidence and limits what can be established.

Evidence verdict

Documented context: Strong support for location, association, and deposition.

Documented ownership history: Stronger support for authenticity and legal-ethical evaluation.

Incomplete provenance: Increased uncertainty and forgery risk.

Invented provenance: Evidence of deception and severe damage to credibility.

08 · COURSE METHOD

Museum numbers and catalog records

Museum and corpus identifiers are essential.

Examples include:

• museum numbers;
• excavation numbers;
• accession numbers;
• CDLI P-numbers;
• inventory numbers;
• manuscript sigla;
• site and locus numbers.

An identifier allows readers to locate the exact object rather than relying on a title such as “the flood tablet.”

A museum record may document:

• object identity;
• dimensions;
• material;
• acquisition;
• excavation;
• findspot;
• date range;
• bibliography;
• images;
• and current interpretation.

A catalog entry is itself a scholarly record subject to revision. Dates and identifications can change.

Archive citation standard

Every featured object should include, where available:

• institution;
• object number;
• title or description;
• material;
• date;
• findspot;
• acquisition history;
• current source;
• and uncertainty note.

11 · COURSE METHOD

Translation is evidence processing

Translation is not the replacement of one word with another.

Ancient languages contain words with multiple meanings, grammatical forms without direct English equivalents, damaged passages, puns, technical vocabulary, cultural categories, and uncertain syntax.

Literal and readable translations

A literal translation preserves structure but may sound unnatural.

A literary translation communicates effect but may conceal ambiguity.

Neither is automatically superior. The research question determines which is more useful.

Translation comparison workflow

When a claim depends on one translated word:

1. Check the original term.
2. Compare several reputable translations.
3. Read the translator’s notes.
4. Determine whether the word is preserved or restored.
5. Ask whether the interpretation depends on only one possible meaning.
6. Avoid building a large theory on an English synonym.

Common failure

A presenter quotes one English translation, treats its wording as exact ancient terminology, and builds a modern technical claim from it.

For example:

• an English version says “flying chamber”;
• the presenter assumes engineered aircraft;
• the original may permit divine vehicle, celestial object, shrine, or metaphor;
• alternative translations are ignored.

Archive standard

The more extraordinary the conclusion drawn from one word, the more urgently the original language and textual condition must be checked.

12 · COURSE METHOD

Textual witnesses and copies

An ancient composition may survive only in later copies.

That does not make it worthless. It changes the question.

Copy date versus composition date

A manuscript has a production date.

The composition may be older.

Evidence for earlier composition can include:

• older fragments;
• quotations;
• linguistic archaism;
• catalogs;
• adaptation by another work;
• historical setting;
• and scribal tradition.

Copying introduces change

Scribes may preserve, modernize, correct, misunderstand, omit, expand, combine, adapt, or comment.

A copy is both:

• evidence for the earlier tradition it transmits;
• and evidence for the later culture that copied it.

Earliest surviving is not necessarily original

The earliest surviving object is the oldest manuscript currently known.

It may not be the first composition.

A later copy can sometimes preserve an older reading than an earlier fragment.

Evidence verdict

Secure: The manuscript existed when the object was copied.

Reasoned inference: The composition or passage existed earlier.

Open question: The exact date and form of lost earlier stages.

13 · COURSE METHOD

Genre controls interpretation

A text’s genre affects what it is trying to do.

Ancient genres include:

• royal inscription;
• hymn;
• lament;
• myth;
• omen;
• law collection;
• treaty;
• letter;
• administrative account;
• genealogy;
• epic;
• ritual instruction;
• wisdom saying;
• king list;
• medical text;
• prophecy;
• satire;
• school exercise.

A royal inscription saying “I destroyed all my enemies” may use conventional victory language.

An administrative ration list has a different relation to daily activity.

A myth may reveal theology and political imagination without functioning as a campaign report.

A law collection may express legal ideals without proving that every rule was uniformly enforced.

Genre error

A genre error occurs when a source is read as though it belonged to another category.

Examples:

• poetry read as technical specification;
• myth read as direct laboratory report;
• propaganda read as neutral journalism;
• ritual text read as proof that everyone believed the same thing;
• genealogy read as a modern census.

Archive standard

Before asking whether a text is factual, ask what kind of text it is.

14 · COURSE METHOD

Testimony and eyewitnesses

Eyewitness testimony can be important.

It is not automatically decisive.

Evaluate testimony through:

• opportunity to observe;
• physical distance;
• timing;
• expertise;
• memory;
• stress;
• motive;
• audience;
• consistency;
• independence;
• and corroboration.

Firsthand but limited

A witness may accurately report what was seen while misunderstanding its cause.

A soldier may describe a bright object in the sky without knowing whether it was astronomical, technological, atmospheric, or psychological.

Later testimony

A memoir recorded decades later can remain valuable.

Time may introduce memory change, narrative smoothing, outside influence, identity formation, and retrospective interpretation.

Anonymous testimony

An anonymous claim may deserve investigation.

Its evidentiary weight is reduced when researchers cannot examine identity, access, conflicts, prior statements, or supporting records.

Multiple witnesses

Several witnesses are stronger only when they are genuinely independent.

Ten websites repeating one anonymous post are not ten sources.

15 · COURSE METHOD

Independence and circular evidence

Independent evidence

Evidence is independent when its informational content does not derive from the same underlying source.

An inscription, a separate archive, an archaeological layer, and an external ruler’s record may converge independently.

Dependent evidence

Three books may all cite one speculative nineteenth-century translation.

They are three publications but one evidentiary foundation.

Circular evidence

Circular reasoning occurs when:

1. a theory is used to interpret ambiguous evidence;
2. the interpretation is then cited as proof of the theory.

Example:

• Assume a symbol is a spacecraft.
• Use that assumption to identify a nearby structure as a spaceport.
• Cite the “spaceport” as proof that the symbol is a spacecraft.

Archive standard

Count independent lines of evidence, not repetitions of the same claim.

16 · COURSE METHOD

Corroboration, convergence, and triangulation

A conclusion becomes stronger when different forms of evidence point in the same direction.

Example of convergence

A city’s destruction may be supported by:

• a dated destruction layer;
• weapon remains;
• an external inscription;
• local administrative interruption;
• abandonment patterns;
• and environmental data.

No single item carries the whole conclusion.

Apparent convergence

Evidence can look independent while sharing one hidden assumption.

Examples:

• several dates calibrated by the same disputed chronology;
• several translations derived from one edition;
• documentaries repeating one author;
• objects from one dealer with invented provenances.

Definitions

Corroboration: One source confirms or strengthens another.

Convergence: Different evidence supports the same conclusion.

Triangulation: Researchers deliberately compare methods or perspectives.

Useful combinations include:

• text and archaeology;
• inscription and administrative record;
• material analysis and provenance;
• linguistic dating and stratigraphy;
• testimony and contemporaneous documentation.

17 · COURSE METHOD

Negative evidence

Absence of evidence

No supporting evidence has been found.

This may reflect poor preservation, limited excavation, wrong search location, inaccessible records, ordinary archival loss, or a false claim.

Evidence of absence

Expected evidence is missing where it should probably survive or be detected.

Negative evidence becomes stronger when:

• preservation conditions are favorable;
• the proposed phenomenon should be widespread;
• the search has been extensive;
• the expected signature is specific;
• and comparable events leave detectable traces.

The absence of one ruler’s name from one damaged tablet is weak negative evidence.

The absence of expected industrial waste, tools, infrastructure, residues, and settlement patterns across well-investigated contexts is stronger evidence against a proposed industrial civilization.

Archive standard

Ask not only whether evidence is absent, but whether the theory predicts that it should be present.

18 · COURSE METHOD

Extraordinary claims

The statement “extraordinary claims require extraordinary evidence” does not mean unusual ideas are forbidden.

It means the quality and quantity of evidence should be proportionate to how much established knowledge the claim would overturn.

Ordinary claim

A newly found tablet contains another copy of a known hymn.

Required evidence:

• object authentication;
• provenance;
• paleography;
• publication;
• comparison with known witnesses.

Extraordinary claim

A tablet proves that extraterrestrials engineered humanity.

Required evidence would include:

• secure object and inscription;
• unambiguous language;
• verified translation;
• clear biological or technological description;
• independent archaeological evidence;
• material traces;
• chronology;
• and exclusion of ordinary literary explanations.

A resemblance to modern science is not enough.

Prior knowledge

Researchers consider what is already known.

This is not closed-mindedness by itself.

It becomes dogmatism only when contrary evidence is dismissed without examination.

A responsible position can be:

• the claim is initially improbable;
• the evidence will still be examined;
• confidence will change if the evidence is strong enough.

19 · COURSE METHOD

Authenticity and forgery

A forgery can imitate material, style, script, damage, patina, provenance, and scholarly vocabulary.

Authentication methods

Depending on the object, researchers may examine:

• material composition;
• tool marks;
• manufacturing process;
• weathering;
• ink;
• pigment;
• thermoluminescence;
• radiocarbon measurements;
• handwriting;
• linguistic anachronisms;
• textual dependence;
• provenance;
• microscopic deposits.

No single method works for every object.

Anachronism

An anachronism places something in the wrong period.

Examples:

• a word not used until later;
• a modern chemical pigment;
• a letter form from the wrong century;
• a quotation from a text not yet composed;
• a modern tool mark.

Genuine material, false inscription

Forgers can use ancient materials.

A test dating the material may not date the inscription or alteration.

Archive standard

A sensational object should become more transparent under scrutiny, not less available for examination.

20 · COURSE METHOD

Peer review, replication, and consensus

Peer review

Peer review means specialists evaluate work before publication.

It can identify missing evidence, methodological errors, unsupported claims, and ignored scholarship.

Peer review does not guarantee truth.

Replication

Replication asks whether observations, measurements, readings, or analyses can be checked again.

Exact replication may be impossible after excavation because excavation changes or removes context. That makes field documentation essential.

Scholarly consensus

Consensus is the current judgment shared by a substantial body of qualified researchers.

It may be based on:

• independent studies;
• stable methods;
• explanatory success;
• repeated criticism;
• and lack of a stronger alternative.

Consensus is not a vote that creates reality.

It is evidence about how specialists assess the record.

Minority interpretation

A minority view is not false because fewer researchers hold it.

It carries a greater burden to explain:

• why the consensus is wrong;
• which evidence is misread;
• what the alternative explains better;
• and how the proposal can be tested.

Archive standard

Do not confuse “not universally accepted” with “equally supported.”

21 · COURSE METHOD

Authority and expertise

Credentials are not evidence for an ancient event.

They are evidence about a person’s likely training and ability to evaluate relevant material.

Relevant expertise

Expertise should match the claim.

Examples:

• Assyriologist for Akkadian grammar;
• archaeologist for excavation context;
• geologist for sediment and catastrophe;
• geneticist for biological ancestry;
• astronomer for orbital claims;
• conservator for object alteration.

A physicist is not automatically an expert in cuneiform.

An archaeologist is not automatically an expert in genetics.

Two authority errors

Weak argument:

A professor said it, therefore it is true.

Also weak:

Experts have been wrong before, therefore expertise has no value.

The correct approach examines both:

• relevant competence;
• evidence and reasoning.

22 · COURSE METHOD

Bias and standpoint

Every source has a standpoint.

Bias does not make a source useless.

It tells the researcher what questions to ask.

Ancient bias

• royal self-presentation;
• temple interests;
• urban rivalry;
• gender hierarchy;
• elite literacy;
• imperial ideology;
• scribal education.

Modern bias

• nationalism;
• religious apologetics;
• anti-religious polemic;
• colonial collecting;
• market incentives;
• academic prestige;
• entertainment value;
• algorithmic attention;
• political identity.

Bias analysis

Ask:

• Who benefits from this account?
• What audience was intended?
• What is emphasized?
• What is omitted?
• What alternatives were available?
• What risks did the author face?
• What genre conventions shaped expression?

Archive standard

Bias is not a magic word that permits dismissal. It is a feature to analyze.

23 · COURSE METHOD

Selection and survival bias

The ancient record is not a neutral sample of ancient life.

What survives favors materials and institutions capable of preservation.

Survival factors

• clay survives differently from papyrus, wood, leather, or textiles;
• stone monuments survive differently from wooden buildings;
• dry environments preserve organic material better;
• elite archives are more likely to be excavated and published;
• cities leave concentrated remains;
• official texts were copied more often than ordinary speech.

Modern selection factors

Collections also reflect:

• excavation priorities;
• colonial interests;
• market demand;
• museum acquisition;
• political borders;
• funding;
• specialist language training;
• publication choices.

The absence of ordinary voices may reflect preservation and power rather than historical insignificance.

Archive standard

The archive is the result of what existed, survived, was found, was collected, and was published.

24 · COURSE METHOD

Digital images and online evidence

Online access expands research.

It also creates new risks.

An image may be:

• cropped;
• mirrored;
• sharpened;
• color-adjusted;
• compressed;
• mislabeled;
• detached from scale;
• taken from a reconstruction;
• or synthetically generated.

Reverse context

A genuine image can be assigned a false story.

The object exists, but the caption does not.

Three-dimensional objects

One photograph can make:

• a broken edge look intentional;
• a shallow mark look engraved;
• a repair look ancient;
• or an ordinary form look anomalous.

Digital verification checklist

• Locate the institutional record.
• Match the object number.
• Compare multiple views.
• Check dimensions and material.
• Find the earliest publication.
• Determine whether the image shows an original, cast, drawing, reconstruction, or artwork.
• Preserve the source and access date.
• Do not use visual resemblance as the only evidence.

25 · COURSE METHOD

Artificial intelligence as a research tool

AI systems can assist with:

• search;
• organization;
• comparison;
• draft summaries;
• terminology;
• and pattern detection.

They can also produce:

• fabricated citations;
• invented quotations;
• blended sources;
• outdated facts;
• false confidence;
• and plausible but unsupported reconstructions.

Safe use

Treat AI output as:

• a lead;
• a draft;
• a query generator;
• or a proposed interpretation.

Do not treat it as the primary evidence.

Verification standard

For factual claims, return to the object record, original text, institutional corpus, publication, or dataset.

For quotations, verify exact wording and location.

For citations, verify that the source exists and supports the statement.

Archive standard

AI may accelerate the path to evidence. It does not replace the evidence chain.

33 · COURSE METHOD

Common evidence failures

“It looks like”

Visual resemblance is treated as identification.

“Ancient people could not have”

The argument begins by underestimating ancient skill.

“No one can explain it”

A presenter has not found or disclosed existing explanations.

“Scientists admit”

No exact scientist, source, or statement is supplied.

“The text clearly says”

The wording appears in only one translation or restored passage.

“There are hundreds of sources”

The sources repeat one original claim.

“It was found near”

Spatial proximity is treated as direct association without stratigraphy.

“The museum dated it”

The date may be broad, stylistic, provisional, or outdated.

“It is too precise to be accidental”

No probability model or comparison set is provided.

“Absence proves destruction”

Preservation and predicted evidence are not evaluated.

“Mainstream scholars refuse to look”

No documented submission, review, response, or test is presented.

“The alternative is possible”

Possibility is treated as probability or fact.

34 · COURSE METHOD

Responsible disagreement

Researchers can examine the same evidence and disagree because of different dating models, language readings, assumptions, evidence weighting, new discoveries, or genuine ambiguity.

Fair representation

Before criticizing a position:

1. State it in a form its strongest advocate would recognize.
2. Identify the evidence it uses.
3. Separate strong points from weak points.
4. Avoid attacking motives unless motive is relevant and documented.
5. Explain why another interpretation is stronger.

False balance

Fairness does not require equal evidentiary weight.

A claim supported by one mistranslated line should not be presented as equal to a conclusion supported by many independent sources.

Archive standard

Respect people equally; evaluate claims unequally according to evidence.

35 · COURSE METHOD

Uncertainty as a result

Historical research often ends with:

• probable;
• disputed;
• unknown;
• text broken;
• context lost;
• date uncertain;
• identification possible;
• authenticity unresolved.

This is not weakness.

False precision is weaker.

A good uncertainty statement explains:

• what is known;
• what is unknown;
• why the issue cannot be resolved;
• and what evidence could help.

Bad:

Scholars have no idea what this symbol means.

Better:

The symbol occurs in three known contexts. A celestial interpretation is plausible because of its position near astral signs, but the corpus remains too small for a secure identification.