Bag of Words
Representing a document as counts over a fixed vocabulary, discarding word order entirely.
One column per token, one row per document, each cell a count. "The dog bit the man" and "the man bit the dog" produce identical rows, which tells you exactly what was thrown away.
It still earns its place, because a bag-of-words or TF-IDF baseline trains in seconds, explains itself token by token, and beats a badly-tuned embedding pipeline more often than write-ups admit. N-grams buy back local order — not good becomes its own column — at the cost of a vocabulary that can grow from tens of thousands to millions.
Three things it cannot do: car and automobile are orthogonal columns so there's no notion of synonymy, order beyond the n-gram window is invisible, and a token absent from the training vocabulary is silently dropped.