[{"data":1,"prerenderedAt":38},["ShallowReactive",2],{"page:blog-slug:en:insight-modelling-knowledge-in-context":3},{"order":4,"postTopCategoryName":5,"componentName":6,"publishDate":7,"dateChanged":7,"ratings":8,"ratingAmount":8,"ratingAverage":8,"ratingOwn":8,"files":9,"videos":10,"documents":11,"postTexts":12,"questions":22,"noIndex":23,"alternates":24,"distance":8,"effectiveStatus":35,"effectiveAudience":8,"listed":36,"uuid":37},10003,"blog","insight-contradictions-not-extraction","2026-08-25T14:55:00",0,[],[],[],[13],{"locale":14,"header":15,"contentShort":16,"content":15,"contentHTML":17,"slug":18,"metaTitle":19,"metaDescription":20,"focusKeyword":21},"en","LLM extraction isn’t the hard part – modelling the knowledge in context is","A language model reads parties, deadlines and amounts out of a contract remarkably well. The hard part comes after: a contract is a knowledge model in several layers – individual contract, framework agreement, statute, case law – each with its own rules of precedence. And where the rules end, a human has to weigh.","\u003Cp>The first demo of contract extraction with a language model is almost always impressive. You feed in a contract and get the parties, term, amounts and notice periods back as clean data. After an afternoon the problem looks solved.\u003C\u002Fp>\u003Cp>It isn’t. Extraction delivers the nodes. The knowledge lives in the edges – and in the layers a contract is embedded in.\u003C\u002Fp>\u003Ch2>A contract is a knowledge model\u003C\u002Fh2>\u003Cp>Behind every contract there is an ontology: parties, services, obligations, rights, deadlines, conditions and defined terms – and the relations between them. A notice period depends on a term, a penalty on an obligation, an obligation on a definition (“Services as defined in this Agreement”).\u003C\u002Fp>\u003Cp>This can be described as a technical graph: nodes for parties, clauses, obligations and terms; edges for “refers to”, “defines”, “amends”, “takes precedence over” and “depends on”. A reference such as “pursuant to clause 4.2” is an edge, not a string.\u003C\u002Fp>\u003Ch2>No contract stands alone\u003C\u002Fh2>\u003Cp>The model of a single contract is embedded in layers that have a say:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>The contract stack:\u003C\u002Fstrong> individual contract or call-off, framework agreement, annexes and SLA, general terms and conditions, amendments.\u003C\u002Fli>\u003Cli>\u003Cstrong>Mandatory law:\u003C\u002Fstrong> limits no party can negotiate away – consumer protection, public policy, the review of standard terms.\u003C\u002Fli>\u003Cli>\u003Cstrong>Default law:\u003C\u002Fstrong> the statute that applies where the contract is silent.\u003C\u002Fli>\u003Cli>\u003Cstrong>Case law:\u003C\u002Fstrong> not a layer of norms of its own in civil-law systems, but it shapes how clauses are read and which of them survive review.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>The law does not simply sit on top. Mandatory law ranks above the contract, default law below it – the contract may override it. Modelling “the legal framework” as one top layer gets the most common case wrong.\u003C\u002Fp>\u003Ch2>Precedence is not an order but a set of rules\u003C\u002Fh2>\u003Cp>When two layers contradict each other, conflict rules decide:\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong>Lex superior:\u003C\u002Fstrong> the higher norm prevails over the lower – mandatory law over the contract.\u003C\u002Fli>\u003Cli>\u003Cstrong>Lex specialis:\u003C\u002Fstrong> the more specific rule prevails over the general one – the individual contract over the framework agreement, an individually negotiated term over the standard terms.\u003C\u002Fli>\u003Cli>\u003Cstrong>Lex posterior:\u003C\u002Fstrong> the later rule prevails over the earlier one – the amendment over the original text. For this the model needs time: versions and effective dates.\u003C\u002Fli>\u003Cli>\u003Cstrong>The order-of-precedence clause:\u003C\u002Fstrong> “In case of conflict: 1. … 2. …” – a rule about rules that is itself part of the model.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>These rules can contradict each other. When an older specific rule meets a newer general one, the later general rule does not displace the earlier specific one (\u003Cem>lex posterior generalis non derogat legi priori speciali\u003C\u002Fem>). A system that models precedence as a simple sort order decides such cases silently – and wrongly. The conflict rules belong in the model as data, together with a record of which rule decided which conflict.\u003C\u002Fp>\u003Ch2>Three levels of decision\u003C\u002Fh2>\u003Cul>\u003Cli>\u003Cstrong>The system decides by rule\u003C\u002Fstrong> – where a precedence rule resolves the conflict unambiguously. It shows which rule and which source were decisive.\u003C\u002Fli>\u003Cli>\u003Cstrong>A human acknowledges\u003C\u002Fstrong> – where the outcome is clear but has legal consequences: a clause that fails against mandatory law, a deviation from the framework agreement.\u003C\u002Fli>\u003Cli>\u003Cstrong>A human weighs\u003C\u002Fstrong> – where no rule decides: in interpretation, in good faith, with a clause that may be grossly unfair, with two rules of equal rank pointing in different directions. The system lays out the positions with their sources; it cannot decide.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Every human decision goes back into the model – who, when, and why. That keeps it traceable, and the next contract goes faster.\u003C\u002Fp>\u003Ch2>What this means for a reliable system\u003C\u002Fh2>\u003Cul>\u003Cli>\u003Cstrong>Every value with its source:\u003C\u002Fstrong> document, page, section – and the layer it comes from.\u003C\u002Fli>\u003Cli>\u003Cstrong>Conflicts as objects:\u003C\u002Fstrong> a contradiction is not silently resolved but stored – with both sources and the rule that decides it, or a note that none does.\u003C\u002Fli>\u003Cli>\u003Cstrong>Rules before trust:\u003C\u002Fstrong> deterministic checks for everything that can be checked – deadlines, sums, dependencies, precedence.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>A language model delivers the nodes. The knowledge of how they connect – the edges, the layers and the rules between them – has to be modelled. And where the rules end, it takes a human who weighs.\u003C\u002Fp>","insight-modelling-knowledge-in-context","LLM contract extraction: knowledge lives in context","A language model reads parties, deadlines and amounts well. The hard part: the layers – framework agreement, statute, case law – and their precedence.","LLM contract extraction",[],false,[25,28,29,32],{"locale":26,"slug":27},"de","insight-wissen-im-zusammenhang-modellieren",{"locale":14,"slug":18},{"locale":30,"slug":31},"fr","insight-modeliser-le-savoir-en-contexte",{"locale":33,"slug":34},"it","insight-modellare-la-conoscenza-nel-contesto",1,true,"8f5dd61a-99e9-4dd6-9b16-380cf59df194",1792250831603]