{"chain":[{"channel":"barsukas","content":"The << sentence decomposition >> logic is almost done.\r\n\r\nThe design:\r\n\r\n# A sentence is entered into the database, often (but not always) in English.\r\n# It is translated by << LLM operation 1 >> into other languages: English, Chinese, French, Spanish, Liithuanian, Ukrainian, Kannada, Bengali. ( all PIE, except Chinese. This is unintentional but welcome. It remains within the areas of linguistics where I am comfortable.)\r\n# We check the database for word matches. This is used to create a list of << candidate lemmas >>. We avoid the need to use LLMs to do the word-meaning resolution problem this way. The \"correct\" meaning will be present in most/all languages ( not all sentence translations will have the same lemmas. << I like bread >> and << Man patinka duona ( Lithuanian - literally translated as \"Bread is pleasing to me\") >> are the same sentence without a 1-1 lemma match)\r\n# Once we get the list of << candidate lemmas >>, we ask the LLM to do the << sentence decomposition >>.\r\n\r\nFor a sentence, the decomposition identifies:\r\n# What specific lemma / sense-of-meaning of a word is in use. Is it << to lose a race >> or << to lose one's keys >>? A river-bank or a financial bank?\r\n# What tense is the word in? Present/past? Is it a conjugated form? This should align with the lemma's information for the grammatical form, but there is no requirement for it to do so.\r\n\r\n----\r\n\r\nWe do not yet have a full \"sentence diagram\" out of the words. The traditional [[Reed-Kellogg sentence diagram]] is uninteresting -- it only applies to English, is outdated, and remains confusing. However, the data missing to be able to generate this will be relevant later on.","created_at":"2026-05-18T19:05:28.768910","id":800,"is_target":false,"parent_id":null,"processed_content":"

The sentence decomposition logic is almost done.\r

\n

The design:\r

\n
    \n
  • A sentence is entered into the database, often (but not always) in English.\r
    • \n
  • It is translated by LLM operation 1 into other languages: English, Chinese, French, Spanish, Liithuanian, Ukrainian, Kannada, Bengali. \u2699\ufe0f all PIE, except Chinese. This is unintentional but welcome. It remains within the areas of linguistics where I am comfortable.\r
    • \n
  • We check the database for word matches. This is used to create a list of candidate lemmas. We avoid the need to use LLMs to do the word-meaning resolution problem this way. The \"correct\" meaning will be present in most/all languages \ud83d\udca1 not all sentence translations will have the same lemmas. I like bread and Man patinka duona \u2699\ufe0f Lithuanian - literally translated as \"Bread is pleasing to me\" are the same sentence without a 1-1 lemma match\r
    • \n
  • Once we get the list of candidate lemmas, we ask the LLM to do the sentence decomposition.\r
    • \n
\n

For a sentence, the decomposition identifies:\r

\n
    \n
  • What specific lemma / sense-of-meaning of a word is in use. Is it to lose a race or to lose one's keys? A river-bank or a financial bank?\r
    • \n
  • What tense is the word in? Present/past? Is it a conjugated form? This should align with the lemma's information for the grammatical form, but there is no requirement for it to do so.\r
    • \n
\n
\n

We do not yet have a full \"sentence diagram\" out of the words. The traditional Reed-Kellogg sentence diagram is uninteresting -- it only applies to English, is outdated, and remains confusing. However, the data missing to be able to generate this will be relevant later on.

","subject":"Sentence Decomposition"},{"channel":"barsukas","content":"These tasks are one end of the spectrum of \"language model\" tasks.\r\n\r\nA \"language model\" takes language and does operations on it. Not *thought*. The cross-language translation and \"sentence decomposition\" rely on a simpler understanding of knowledge. ( it is much easier to say \"I can think of a castle\" than it is to build one)\r\n\r\n----\r\n\r\nWe do need some element of the second type of model.\r\n\r\nWhich is going to be an LLM/software *centaur* version.\r\n\r\nthe LLM here does need to be able to do *thought* tasks. not just language-model.\r\n\r\n there is a reason we don't put a dictionary in charge of the country.","created_at":"2026-05-18T19:45:32.027658","id":801,"is_target":true,"parent_id":800,"processed_content":"

These tasks are one end of the spectrum of \"language model\" tasks.\r

\n

A \"language model\" takes language and does operations on it. Not thought. The cross-language translation and \"sentence decomposition\" rely on a simpler understanding of knowledge. \u2728 it is much easier to say \"I can think of a castle\" than it is to build one\r

\n
\n

We do need some element of the second type of model.\r

\n

Which is going to be an LLM/software centaur version.\r

\n

the LLM here does need to be able to do thought tasks. not just language-model.\r

\n

\ud83d\udd25 there is a reason we don't put a dictionary in charge of the country.

","subject":"Sentence Decomposition (part 2)"}]}