| I'm not saying that this is implausible, but this is one guess in a sea of many. We were shown a single word substitution with a claim that 300 others match, but no documentation of this person's research. We don't know this person's methodology beyond this single sentence. A naive approach would be to collect a thesaurus of Semitic roots and use automation such as an LLM to match those against instances of words. Sounds plausible in the beginning, but words do not exist independently. To illustrate how things can go wrong, let's try to prove that English is a Semitic language. Suppose that the source material we have is this sentence: "A baker brought a bushel of wheat to the mill." Now let's match it to Semitic roots as found on this list: https://en.wiktionary.org/wiki/Appendix:Proto-Semitic_stems. Let's look for plausible substitutions now: - "Baker" matches "bVkVr-", "first-born son". We have our anchor now. (V means any vowel). - "Brought" looks plausibly close to "burāṯ-", "juniper" on our list. So far so good. - "Bushel" is a good match with "b-š-l", "to be cooked"! - "Wheat" does not have an exact match. It could be a loanword, for example. - "Mill" looks like "m-r-r", "bitterness" if we assume lack of written L/R distinction. Again, juniper + cook + bitter is plausible, because juniper can be bitter. - The meaning of particles will be inferred from the sentence structure. Okay, let's take a look at our translation! We have "first-born son", "juniper", "cook", (wheat), "bitter". Pretty clear that (wheat) must be the name of a dish here. Therefore, the sentence can be translated as "A bitter juniper dish is being cooked for a first-born son". This even matches the context: the sentence was found in a granary, and it refers to food. My point here is that with such a small sample size, we can extrapolate the data to mean absolutely anything. With no reference material, we cannot assess the correctness of any translation. |