llms-ppi.txt

1. The Tripartite Structure of the PPI

According to the JSON architecture, the PPI is organised into three main sections:

• Introduction (static/instructive): Describes the core question, the role of the AI, and the template. This is where the so-called "screening process" is anchored, which governs the fundamental comprehension of the text.
• Namespaces (dynamic/generated): This is the heart of the PPI. Here, various philosophical categories of analysis are defined, interconnected through logical relations and dependencies.
• Usage (static/instructive): Defines the concrete workflow, best practices, and documentation templates for the results.

2. The Central Namespaces (Categories of Analysis)

The namespaces within the PPI correspond to the perspective tags of the PTL (such as <BEGRIFFSANALYSE /> or <ONTOLOGISCHEANALYSE />) and each assign the AI a quite specific philosophical role. Among the most important namespaces of the Philosophical Turn are:

• PhilosophicalTurn.PerspektivitaetsAnalyse: Disentangles the potential for ambiguities in concepts such as "subjective/objective" and "abstract/concrete".
• PhilosophicalTurn.Topologie: Maps out the logical and discursive field of a text in order to render categorial boundaries visible.
• PhilosophicalTurn.Alethologie: Analyses the concepts of truth employed in the text and their functional role within the argumentation.
• PhilosophicalTurn.Begriffsanalyse: Operationalises conceptual work as the controlled clarification of meaning and delimitation – the concern here is not with things, but with linguistic tools.
• PhilosophicalTurn.OntologischeAnalyse: Identifies fundamental metaphysical assumptions (What exists? What is the nature of entities and properties?).
• PhilosophicalTurn.ErkenntnistheoretischeAnalyse: Reconstructs a text's concept of knowledge and its architecture of justification, examines chains of reasoning with respect to the distinction between demonstrative and problematic induction, and surveys the limits of cognition through the categorial distinction between a soluble problem and a mystery in principle, in order to expose the actual tenability of epistemic claims.
• PhilosophicalTurn.EpistemischeAnalyse: Identifies and reconstructs the concept of knowledge, the structures of justification (chains of reasoning, evidential basis), as well as the sources and limits of cognition, in order to reveal the discrepancy between a text's claim to justification and its actual justificatory achievement.
• PhilosophicalTurn.GeltungstheoretischeAnalyse: Examines purely the basis of legitimation and the architecture of argumentation, that is, how and on what grounds claims to validity are justified.
• PhilosophicalTurn.NormativeAnalyse & AxiologieAnalyse: Reconstructs norms, duties, and rights in strict distinction from descriptive statements, and examines dimensions of value (intrinsic vs. instrumental).
• PhilosophicalTurn.DeontologischeAnalyse: Systematically unfolds the family of deontological approaches by way of the guiding distinction between universal and contingent deontology, and analyses moral judgements with respect to their inherent fittingness as well as their stability in terms of validity theory (viscosity).
• PhilosophicalTurn.InklusionsAnalyse: Deconstructs social-ontological theories of inclusion and exclusion along three axes (directional diagnostics, contingent deontology, status vs. existence), in order to uncover tipping points, drift effects, and concealed forms of exclusion (such as "soft heteronomy").
• PhilosophicalTurn.PluralistischeAnthropologie: Reconstructs the anthropological deep structure of a text in order to reveal how implicit conceptions of the human being (as ostensibly descriptive premises) function as a concealed "engine of deduction" for normative, institutional, and political conclusions.
• PhilosophicalTurn.EigentumsAnalyse: Deconstructs argumentations grounded in property theory and reconstructs the normative, legal, and social-ontological premises through which possession, rights of disposal, and claims to property are legitimated, derived, or criticised.
• PhilosophicalTurn.HabermasAnalyse: Examines texts with regard to their discourse-ethical and communication-theoretical presuppositions, analyses the redeemability of claims to validity (such as truth, rightness, sincerity) within communicative situations, and examines the field of tension between discursive normativity and strategic action.

3. Usage

The use of the PPI follows a strict, methodical procedure ("Usage"), which ensures that the AI's analyses remain profound and comprehensible:

• Step 1: Understanding the Text and the Task (Screening Process): Before specific analyses are applied, the text must be read and its principal philosophical characteristics identified.
• Step 2: Logical Order of Application: The namespaces may not be invoked arbitrarily. They must be arranged in a logical sequence, whereby prerequisites ("requires" dependencies) are to be addressed first.
• Best Practices: A direct leap to the namespaces without prior systematic screening is to be avoided. Furthermore, the PPI requires that all claims made by the AI be substantiated with concrete textual references and that cross-connections between the various analyses (namespaces) be made evident.
• Documentation Template: The AI is instructed to structure the results cleanly. The output should contain the screening result, the specific namespace analyses, a critical evaluation (freedom from contradiction, transparency), as well as a list of ambiguities and limitations of the analysis.

In summary: The PPI is the conceptual "library" of philosophical perspectives that one draws upon when writing a PTL prompt. By embedding the PPI namespaces as tags (e.g., <NORMATIVEANALYSE />) within a PTL prompt, one guides the cognitive methodology of the AI in precisely the manner that a lecturer in a philosophical seminar guides the mode of thought of his students.

Philosophical competence explicitly does not consist in the mere accumulation of knowledge or the reproduction of memorised facts – not even in facts concerning authors of the philosophical canon. A competence-oriented approach draws a sharp dividing line between „know-that“ (pure factual knowledge) and „know-how“ (the methodical capacity for thinking). True scholarship and philosophical competence therefore mean generating one's own knowledge rather than merely reproducing that of others.

1. The „dramaturgy of thinking“ and self-directed activity 

In examinations and written assignments, competence does not manifest itself in the passive rehearsal of material, but in the „dramaturgy of thinking“. This means possessing the ability to render an intellectual problem methodically discernible, to work it through, and to adopt a well-founded position. What is at stake is the self-directed activity of thought, a profound understanding of the structure of arguments, and the capacity to defend one's own position „clearly, in a differentiated manner, and resiliently“ against objections. The ultimate goal is to internalise the „academic habitus of thinking“.

2. Active engagement with problems rather than passive consumption 

For most people, reading is a passive process of reception. A competent student of philosophy does not wait for predetermined, alternative-less answers from canonical authors or instructors. Whoever merely asks questions surrenders themselves to ostensible authorities. Philosophical competence manifests itself in the principle: „Ask less, act more“. One must be capable of making well-researched proposals and of conceiving of tasks as a field of play for one's own reflection. In doing so, one learns to formulate philosophical questions with precision – questions that are not mere grammatical queries or simple yes/no decisions, but intellectual puzzles and fields of tension that must be unfolded argumentatively.

3. The end of „pseudo-competence“ in the age of AI

The availability of artificial intelligence compels the further development of philosophical competence. Since AI can retrieve factual knowledge and summaries without error, it brings to an end the age of „memorised pseudo-competence“, by means of which one could hitherto muddle one's way through one's studies. In order not to become superfluous, one must acquire a „competence of competence“. Classical pseudo-competence reveals itself above all when professionals can only think and speak in the idiosyncratic „linguaggio“ of their sanctified idol. This is the capacity to actively steer one's own process of cognition by developing precise questioning strategies and by specifying the methodical parameters (such as the definition of namespaces in the PPI) with exactitude. It is a learning process in which one must, figuratively speaking – like Baron Münchhausen – „pull oneself out of the swamp of obscurities by one's own hair“, by using AI as a partner in reflection and a catalyst, without surrendering one's own thinking to it.

4. Categorial precision and methodical disentanglement 

On the level of craftsmanship, philosophical competence consists in the ability to methodically analyse and disentangle texts, rather than reflecting upon them in an unstructured manner. This includes, among other things:

• Recognising the definition-in-use: Analysing what work a concept performs within a concrete argument, rather than drawing on dictionary definitions.
• The strict separation of disciplinary levels: Preventing questions of being (ontology), of knowing (epistemology), of value (axiology), and of obligation (normativity) from being conflated.
• The reflection upon implicit premises: Recognising what a text tacitly presupposes (e.g. ontological commitments or unnamed sources of value), and rendering these blind spots visible.

Ultimately, philosophical competence means enduring and methodically handling the irreducible plurality of philosophy: one learns that there is no objective unified method, and that one must instead always precisely define and reflect upon one's own respective analytical perspective.

Simply providing the llms.txt as a source for the AI chat has disadvantages. If one hands over the entire, highly complex PPI-JSON together with a primary philosophical text to an AI, accompanied by the simple instruction "Analyse this", the AI will inevitably capitulate. On the one hand, it is overtaxed in terms of capacity.  On the other hand, it then falls into what may be called "namespace inflation": it merely skims many concepts superficially instead of proceeding methodically into depth.

The solution to this problem consists in a "graduated prompting" – a procedure that may be compared to the "laminating of puff pastry", in which prompts build upon one another through several foldings. Owing to its architecture (in particular the areas usage and screening_process), the PPI-JSON is already perfectly prepared for precisely this stepwise application.

In order to have the AI not merely "read" the JSON but employ it as a genuine "conceptual corset", you must compel it through the workflow. Here are three practical examples of application showing how you may deploy the PPI in prompt chains (cascades):

Example 1: The Screening Prompt (Delimiting the Problem)

You must not allow the AI to analyse immediately. In the first step, you upload the JSON (e.g. as a file) together with your philosophical text. You instruct the AI to carry out exclusively the screening process, in order to select the appropriate tools.

Practical prompt:

"Attached you will find my PPI framework (JSON) as well as a philosophical text. Read both. Now, strictly in accordance with the specifications under usage.workflow 'Step 1 & 2', carry out a screening of the text. Do not yet apply any analyses! Instead, produce for me exactly the 1. SCREENING REPORT from the documentation_template. Explicitly justify which namespaces from the JSON are relevant for this text, which are to be excluded, and which requires-dependencies must imperatively be observed."

Why this works: The AI is compelled to regard the JSON as a toolbox. It first considers what it is to do, before doing it. As a response, you receive a clear roadmap.

Example 2: The Cascade Prompt (Working Through Dependencies)

Once the screening has determined which namespaces are relevant (e.g. for an epistemological text), you compel the AI to observe the logical order of the JSON (relation_types_guide). The AI must not leap to the goal before the foundational work has been accomplished.

Practical prompt:

"We continue our work on the basis of your screening report. You have determined that we require the PhilosophicalTurn.ErkenntnistheoretischeAnalyse. The PPI stipulates, under the relations, that the PhilosophicalTurn.Begriffsanalyse constitutes a mandatory precondition (requires: mandatory) for this. Now carry out FIRST the conceptual analysis. Clarify the central epistemological termini of the text by reference to the kernfragen and kernmethoden of this particular namespace. Produce the 2. APPLICATION REPORT solely for the conceptual analysis. Stop thereafter. We shall undertake the epistemological evaluation in the next step, on the basis of these clarified concepts."

Why this works: You prevent the typical AI error of employing imprecise everyday concepts for the analysis of argumentation. By halting the AI, you enforce a high degree of "analytical acuity" for the individual step.

Example 3: The Error-Avoidance Prompt (Enforcing Categorial Precision)

When it comes to substantive depth (e.g. in complex ontological or normative questions), AI tends to commit category errors. You can employ the PPI structure to pin the AI down explicitly to the error diagnostics of the respective namespace and to the global meta_hinweise.

Practical prompt (example: free will):

"Now apply the namespace PhilosophicalTurn.OntologischeAnalyse to the text at hand concerning free will. Proceed as a precise analytical philosopher and imperatively observe the specification funktions_und_arbeitsweise from the introduction: never conflate ontological positings with epistemic states. Examine the text specifically by means of the 'Continuity vs. Occurrence Test (Broad)'. In doing so, attend strictly to the typical error named in the meta_hinweise, namely 'Vicious Abstraction' – does the text reify freedom as an independent entity? Document your findings and cross-connections in the format 3. SYNTHESIS."

Why this works: Through targeted references to key terms from the JSON (such as Vicious Abstraction or Continuity Test), you activate precisely those weights within the AI's neural network that pertain to C.D. Broad's method and to rigorous analytical philosophy. In so doing, you fit it with the desired "conceptual corset".

Conclusion for practice: Do not treat the PPI-JSON as a command that the AI is to execute in one single pass, but rather as a set of rules for a role play. You assume the role of the instructor who prescribes the methodology, and you steer the AI's process of cognition through small, verifiable stages (screening -> foundational analysis -> in-depth analysis -> synthesis).