As noted in the comments, there are issues with that Wikipedia article.
As with many confusions in English grammar, this can be traced to people trying to think about English using terms adapted for Latin. In Latin, the subject of a verb can be indicated by the form of the verb (amo = I love, amas = you love etc) So there are some words in English that don't translate to words in Latin, but to elements of Lating morphology. And since, in Latin, morphology is much more important than syntax, such words seem to be "words that do grammar". They don't have any "meaning" (since meaning is understood to be the translation of the word to Latin).
This is all a bit confused, because English isn't Latin...
All words have functions, all words have some content. It may be useful to contrast words like "cat", "eat", "big" with words like "I", "will", "the". The most common terms for these are "content or lexical" words, contrasted with "functional or grammatical" words. But be aware that these labels are just labels. It may also be better to analyse at the morpheme level, rather than at the level of words. All morphemes have functions. Lexical morphemes can give meaning on their own. Functional morphemes are normally attached to lexical ones. New lexical morphemes can be created, ad hoc. New functional morphemes develop much more rarely. But there are grey areas, and words can move from one category to another. For example, "will" is primarily used to indicate a future time. But it still has a lexical sense.
Reading: https://www.thoughtco.com/content-lexical-word-1689918
