Academic and Engineering Walkthrough
Formal logic asks whether the conclusion follows from the premises. Argumentation theory adds: does the move perform the task of this dialogue? Information about the source’s competence is important when evaluating testimony and irrelevant when checking an algebraic proof. Context does not justify everything, but it defines the participants’ obligations.
flowchart TD
A["Реплика"] --> B{"Поддерживает
исходный тезис?"}
B -->|Да| C{"Основание надёжно?"}
B -->|Нет| D{"Меняет тему,
позицию или человека?"}
C -->|Да| E["Релевантный аргумент"]
C -->|Нет| F["Проверить источник
и силу свидетельства"]
D -->|Да| G["Ошибка релевантности"]
D -->|Нет| H["Уточнить цель диалога"]
Chapter Rule
Check not the topic of the replica, but its operation. How exactly does this fact increase
the justification of this thesis?
1. Ad hominem: against a person
1.1. Direct attack
Автор RFC некомпетентен.
Следовательно, предложенная схема идемпотентности неверна.
The author’s character does not indicate a schema error.
1.2. Circumstantial ad hominem
Он предлагает managed database, потому что работает у облачного провайдера.
Следовательно, аргументы о стоимости неверны.
Conflict of interest reduces trust and requires independent verification, but does not
make every statement false.
1.3. Tu quoque: «you are the same»
Ты требуешь code review, но сам вчера смержил без него.
Следовательно, code review не нужен.
A person’s inconsistency can be genuine and does not disprove the rule.
But in the dispute over the fairness of applying the norm this is relevant:
Why is the exception available to the manager and unavailable to the team?
1.4. Source poisoning
The audience is told that anyone who disagrees is selfish, foolish, or dangerous. Thus, potential grounds are blocked regardless of the content.
1.5. Guilt by Association
Этой библиотекой пользуется сомнительная компания.
Следовательно, библиотека небезопасна.
A porting mechanism is needed: a common maintainer, supply chain compromise,
license dependency. Proximity alone is not enough.
2. When personal data is relevant
Personality matters if the thesis depends on:
- witness testimonials;
- expert qualifications;
- conflict of interest;
- forecast accuracy history;
- promise fulfillment;
- distribution of authority and responsibility.
The form should be bounded:
Источник имеет нераскрытый финансовый интерес.
Значит, его свидетельство нельзя считать независимым без проверки.
No:
Источник заинтересован.
Значит, утверждение ложно.
3. Straw man: straw man
A position is replaced by a weaker or extreme one, the replacement is refuted, and victory is declared.
Original thesis:
For financial operations, a separate audit log with restricted access is needed.
Replacement:
The author wants to log everything and store it forever.
Refuting the second position says nothing about the first.
Types of Substitution
- strengthening: “sometimes” turns into “always”;
- weakening: an exact criterion is replaced with a vague one;
- choosing a weak representative: the worst version of a group is attacked;
- quote without a caveat: the context, time, and modality disappear;
- composite straw man: positions of different people are merged into one.
Check
Ask to formulate the position in such a way that its author would agree with the formulation.
4. Steelman can also substitute
Improving position is useful for research, but one cannot claim that the author
said an improved version.
Реальная позиция: ...
Сильнейшая версия, которую имеет смысл проверить: ...
Distinguish between interpretation and quotation.
5. Red herring: Distracting clue
The discussion leads to a bright related topic.
— Есть ли lawful basis для этой обработки?
— Наш продукт очень полезен пользователям и использует современное шифрование.
Usefulness and encryption are important, but they do not establish a legal basis.
Test
After the response, can one still ask the original question without modification? If yes, the response likely did not close it.
6. Whataboutism
Criticism is answered by pointing out another’s fallacy:
— У нас секрет в репозитории.
— А команда B вообще хранит ключ в wiki.
Comparison may be relevant for the risk portfolio priority, but does not negate
the presence of a secret.
Repair:
Yes, we need to revoke our secret. After containment, we’ll compare the team B’s exposure
and set the second risk in the common backlog.
7. Ignoratio elenchi: irrelevant conclusion
They successfully prove the wrong thesis.
Тезис:
Текущий сервис нельзя независимо развернуть.
Ответ:
Микросервисы позволяют командам работать автономно.
The general utility of an architectural style does not prove the readiness of a specific system.
8. Appeal to Authority
Expert testimony — a normal way to learn the world. A fallacy occurs,
when authority is used outside the conditions.
Criteria for Strong Expert Evidence
- Competence area coincides with the thesis.
- The expert indeed has the corresponding qualification.
- The opinion is conveyed accurately and in context.
- Experts in the field have a reasonable consensus or the disagreement is disclosed.
- Conflict of interest is taken into account.
- The claim allows for external verification.
- Authority does not replace primary data when they are available.
False Authority
A well-known CEO discusses clinical effectiveness; a strong backend engineer is declared an expert in interface accessibility without data or practice.
Anonymous Authority
Эксперты говорят...
Исследования доказали...
В индустрии известно...
Without an identifiable source, it is impossible to check the scope, method, and context.
Lone Dissident
The presence of one degree holder disagreeing with a field does not create
a symmetric scientific dispute.
9. Appeal to popularity
Большинство команд использует Kubernetes.
Следовательно, Kubernetes оптимален для нас.
Popularity may be an indication of:
- mature ecosystem;
- availability of specialists;
- reduction of vendor risk.
But does not prove compliance with scale, competencies, and constraints.
Working output:
Popularity reduces the risk of hiring and tooling shortages; other criteria
are checked separately.
10. Appeal to tradition and novelty
Tradition
Мы десять лет деплоим вручную, значит, процесс правильный.
The age of the practice can confirm stability, but also mean
inertness.
Novelty
Новый фреймворк использует signals, значит, он лучше.
New techniques can resolve old limitations; the mechanism and cost must be named.
The Right Question
Which properties were preserved or changed and why are they relevant?
11. Appeal to nature
Это естественный способ организации, значит, он хороший.
«Natural» is ambiguous:
- occurs without intervention;
- is familiar to a person;
- is biologically determined;
- is environmentally friendly;
- simply sounds positive.
Normative evaluation does not follow from origin. Toxic substances are natural,
vaccines and encryption are artificial.
12. Appeals to Emotions
To Fear
If we don’t buy this product, the next leak will destroy the company.
Fear may signal a real risk, but probability, damage,
alternatives and control effectiveness are needed.
On Pity
Do not reject the PR: the author worked all night.
Effort is important for supporting a person, but does not make the code correct.
On Mockery
Only dinosaurs write monoliths.
Reputational cost replaces decision comparison.
To the Lestnaya Self-Image
Real senior engineers do not ask for documentation.
A person is pushed to confirm the status by rejecting a reasonable requirement.
To Anger and Disgust
An emotional reaction may be morally appropriate and yet not establish
facts or causality.
13. Appeal to Force
Архитектура правильная, потому что CTO её утвердил.
A solution may be mandatory due to authority, but does not become
epistemically true.
Split:
Что считается действующим решением?
Какие основания делают решение хорошим?
Sometimes the authority is relevant precisely to the first question.
14. Appeal to Consequences
Если вывод верен, проект придётся остановить.
Это неприемлемо.
Следовательно, вывод неверен.
Undesirable consequences do not change the truth of a descriptive thesis.
But the consequences are relevant to the decision:
Гипотеза вероятна на 60%.
Цена ложного отрицания критична.
Следовательно, действуем до полной уверенности.
This is decision theory, not a proof of truth.
15. Genetic Fallacy
Statements are evaluated only by origin:
Идея пришла из маркетинга, значит, технически бессмысленна.
Origin may hint at interests and process quality, but content
needs to be checked separately.
Do not confuse with data provenance: the dataset source directly affects
the representativeness and legality of its use.
16. Motive Fallacy
Автор хочет повышения, поэтому оценка риска ложна.
A motive can explain why a thesis is promoted, and does not show whether it is true. However, the motive is relevant to trust, conflict disclosure, and independent verification procedure.
17. False Balance and False Equivalence
False balance
Two positions are given equal weight solely because they exist, ignoring
the quality and volume of evidence.
False Equivalence
Two cases declare identical by superficial similarity:
И hotfix, и обход контроля меняют production без полного процесса; значит,
они одинаково рискованны.
Differences are needed: review, time constraint, audit, rollback, scope.
18. Tone policing
Presentation quality is used to avoid answering the content:
Пока вы говорите так эмоционально, обсуждать утечку бессмысленно.
Tone can interfere with team safety and require a separate conversation,
but the actual risk remains. Split into two tracks:
Сначала containment.
Затем нормы взаимодействия.
19. Answer Map
| Stage |
Check question |
| Ad hominem |
How does a person’s property change the truth of a thesis? |
| Straw man |
Is this an exact formulation of the position? |
| Red herring |
Is the original question closed? |
| Whataboutism |
Does someone else’s error cancel the current one or only change the priority? |
| Authority |
Do the competency match, is the source verifiable, is the consensus considered? |
| Popularity |
What does popularity measure in this task? |
| Tradition / novelty |
How is the age of the practice related to the required property? |
| Emotion |
How does emotion change the data or the decision? |
| Force |
Is it about truth or a mandatory decision? |
| Consequences |
Do consequences affect truth or only action? |
20. Honest Response Protocol
- Repeat the thesis in your own words.
- Ask if the formulation is accepted.
- Name the reply that seems irrelevant.
- Explain the missing bridge.
- Acknowledge the separate value of the reply.
- Return the question.
Да, эксплуатационная сложность Kafka важна. Но исходный вопрос — нужна ли нам
повторная доставка. Сложность инструмента не отменяет требования. Давайте
сначала подтвердим требование, затем сравним реализации.
21. Tasks
For each case specify:
- original thesis;
- response function;
- can the response be relevant in another dialogue;
- corrected utterance.
- “This benchmark is incorrect: its author sells the database.”
- “You want to add a mandatory review to all changes.” — “So you want to stop delivery.”
- “Need to remove the public secret.” — “Competitors have bigger leaks.”
- “Why is the forecast not calibrated?” — “The model uses a modern transformer.”
- “All mature companies have implemented platform teams.”
- “This conclusion cannot be correct: then we will have to close the product.”
Sources
What’s next
The next chapter discusses fallacies where the issue is not in the subject of the answer, but in a hidden assumption, standard of proof, or the language itself: circularity, complex question, burden of proof, equivocation, composition, slippery slope, and others.
Presumption and Ambiguity Fallacies