R — Original Text (Reading)
"To fully reflect the whole thing, reflect the essence of the thing, and reflect the internal regularity of the thing, it is necessary to go through the process of thinking, to refine the rich perceptual material by discarding the coarse and selecting the essential, eliminating the false and retaining the true, proceeding from one to another, and from the surface to the interior, and to create a system of concepts and theories, and it is necessary to leap from perceptual knowledge to rational knowledge." — "On Practice" (July 1937)
I — Methodological Skeleton (Interpretation)
- Raw information (user feedback, data reports, market observations) is only "perceptual material" and is not equal to cognition. It must undergo active processing before it can become a basis for decision-making.
- Processing is divided into four steps, none of which can be skipped, and they are in a progressive relationship:
- Discard the coarse and select the essential: filter out noise and retain key signals. Among 1,000 pieces of feedback, perhaps only 50 are valuable information.
- Eliminate the false and retain the true: identify and exclude false/distorted data. Paid commenters' reviews, self-indulgent research, and beautified reports are all false.
- Proceed from one to another: establish connections. The causality and correlations between user feedback and usage data, and between competitor developments and market trends.
- From the surface to the interior: go from surface phenomena deep into underlying patterns. The deeper meaning of "users say they want X" is "users actually need Y".
- The four steps are not completed in one pass, but are iterated repeatedly. Each round deepens cognition by one level.
- Skipping the earlier steps and going straight to a conclusion ("I feel that's just how it is") is subjectivism; stopping after only the first two steps (having clean data but no insight) is empiricism.
A1 — Application in the Book (Past Application)
Application 1: On Protracted War — distilling strategic judgment from a mass of intelligence
- Problem: In the early period of the War of Resistance, information from all sides was chaotic—some said China was certain to perish, some said it could win in three months, and battlefield news was sometimes good and sometimes bad.
- Use of the method: Mao Zedong used four steps to process all the information on both China and Japan: discard the coarse and select the essential (filter out emotional statements and retain hard data) -> eliminate the false and retain the true (exclude false battle reports and exaggerated propaganda) -> proceed from one to another (establish a correlational analysis of both sides' strong/weak, large/small, progressive/regressive, many helpers/few helpers) -> from the surface to the interior (go from the surface comparison of military strength deep into the underlying patterns of the nature of the war, the direction of people's hearts, and the international situation).
- Conclusion: three stages of protracted war, and final victory belongs to China.
- Result: The prediction of On Protracted War was basically verified by the actual course of the eight-year War of Resistance.
Application 2: Reform Our Study — using the comparative method to drive cognitive upgrading
- Problem: Party cadres treated information with a subjectivist attitude—drawing conclusions without doing investigation, and making judgments while cutting off history.
- Use of the method: Mao Zedong proposed that "shi" (facts) = "all things that objectively exist," "shi" (is) = "the internal connections of things," and "qiu" (seek) = "to study." Using the imagery of "reeds on the wall / bamboo shoots in the mountains," he had cadres self-check the quality of their information processing.
- Conclusion: Only information processed through "seeking truth from facts" can serve as a basis for decision-making.
- Result: "Reform Our Study" became a programmatic document of the Yan'an Rectification Movement, driving the whole Party to shift from subjectivism to seeking truth from facts.
A2 — Trigger Scenarios (Future Trigger)
- User feedback flood: After the product launch, a large amount of user feedback is received, with emotional complaints and rational suggestions mixed together. Users will say: "Users say all sorts of things," "There's too much feedback to get through," "I don't know which to listen to and which not to."
- Contradictory data: Different data sources give contradictory signals — sales says the market is very good but retention data is declining; users say they are satisfied but the renewal rate is falling. Users will say: "The data is fighting itself," "I don't know which data to believe."
- Competitor information overload: Competitor developments are frequent and there are too many industry reports; the volume of information is large but actionable conclusions are few. Users will say: "I've read a pile of reports but don't know the conclusion," "Competitor information is too messy."
- Distorted reporting: Subordinates report only good news and not bad, data is beautified, and real problems are covered up. Users will say: "I always feel that what is reported isn't true," "The data looks good but the problems are hidden."
- Insufficient decision information: A judgment must be made when information is incomplete, and reasoning must be done from limited clues. Users will say: "There isn't enough information but a decision must be made," "I can only judge based on these clues."
Language signals: When users mention words such as "too much noise," "signal," "core information," "data fighting itself," "surface and deep," "essence," etc., this skill should be activated.
Distinguish from adjacent skills：
- vs Investigation and Research Method: this skill = processing information (what to do after obtaining raw material), the latter = collecting information (how to obtain good raw material)
- vs Practice-Epistemology: this skill = the technique of a single round of cognitive processing, the latter = the philosophical framework of multiple rounds of cycles
- vs Contradiction Analysis Method: this skill = general steps for handling information, the latter = an analytical framework for finding bottlenecks
E — Executable Steps (Execution)
Step 1: Discard the coarse and select the fine + discard the false and retain the true (information cleaning)
- List all raw information in a source inventory. Annotate each piece of information with: (1) source credibility (2) whether it can be cross-verified (3) whether it is emotional/subjective.
- Filter out: emotional expressions, unverifiable single sources, and obviously beautified/uglified data. Retain: hard data that can be cross-verified, and signals consistently indicated by multiple sources.
- Completion standard: The amount of raw information is reduced to within 20%, and the remaining information all has source annotations and credibility ratings.
Step 2: From this to that (establish connections)
- Classify the cleaned information by dimension (user side/product side/market side/competition side), and look for cross-dimensional connections and causal chains.
- Draw an information relationship diagram: A changes -> B changes -> C changes, and identify the key transmission paths.
- Completion standard: At least 3 cross-dimensional relationships are discovered, and each relationship is supported by at least 2 independent pieces of evidence.
Step 3: From surface to interior (go deep into the essence)
- For each key association, ask "why" at least 3 levels deep: Why does the user do this? Because ___. Why ___? Because ___. Why ___?
- Distinguish surface needs (what the user says) from deep needs (what the user actually needs), and confirm that what you have found is the latter.
- Stopping point: If, by the 3rd level of questioning, it is still only repeating the surface-level "because the user likes/dislikes it," that means the information is not deep enough, and you need to return to step 1 to supplement it with deeper information (user interviews rather than survey data).
- Completion criteria: Produce a "surface-deep" comparison table, and every deep insight must be able to guide a specific action decision.
B — Boundary
Scenarios not to use
- When the information is already clear: If the data is already structured and the causal chain is already clear, there is no need to go through the four-step method again; decide directly.
- When a quick response is needed: In emergencies, use intuition and experience to make immediate judgments, and afterward use the four-step method to review.
- Creative work: In scenarios such as design and creative work that require divergent thinking, the four-step method will suppress creativity.
Failure modes (from counterexamples)
- Party stereotype writing (ce08): Information processing becomes a formalistic performance—reports get longer and longer but have no core conclusion, listing phenomena as A, B, C, D but lacking causal analysis. "Putting on airs to intimidate people"—using an authoritative tone in place of real analysis.
- Empiricism (ce02): Stopping at the "discard the dross and select the essence" stage, satisfied with clean raw data, but not doing the deep thinking of "from one to the other, from the outside to the inside." The result is data without insight.
- Being misled by surface phenomena (ce23): Skipping the "from the outside to the inside" step and treating a temporary phenomenon as a long-term trend. One defeat means "it's all over," and one victory means "unstoppable."
Author's blind spot
- Mao Zedong emphasized "seeking truth from facts," but during the Great Leap Forward he himself believed false per-mu yield data (the "discard the false and retain the true" link failed). Lesson: the information processor's stance and wishes can affect the judgment of "discarding the false"—the more you want to believe information, the more strictly you must verify it.
Related skills
- Practical epistemology (composes-with): The seek-truth-from-facts information processing method is a concrete operational technique for the leap from "perceptual knowledge → rational knowledge" in practical epistemology. Practical epistemology provides the overall framework of the cognitive cycle, and the four-step seek-truth-from-facts method provides the specific steps of information processing within a single cycle.
- Investigation and research method (composes-with): The investigation and research method is responsible for "collecting high-quality first-hand information" (raw material acquisition), and the seek-truth-from-facts information processing method is responsible for "processing raw materials into a basis for decision-making" (raw material processing). First investigate and collect, then process according to seek truth from facts, forming a complete information processing chain.