Vowels are the carriers of speech, consonants the articulations. Or put in different terms, vowels garantee speech flow, consonants the differenciation, the discreteness of language.

Only in the combination of both, however, lies the essence of language and occurs the birth of that hard to define concept, the word.

As C.C. Fries would have it, a word cannot be genuinely defined, only sequences of sounds as uttered during a conversation, the sequences constituted by the turn-taking of talk, only this going to and fro between human beings can be considered a "natural unit" of language. A word, however, cannot be non-trivially derived from a sequence. A word, however, is also indispensable as a fundamental unit for grammar.
This difference in subdivision, or this different approach in anatomy of language is of great consequences. Whereas sequences in a dialogue, e.g., often display skip-connecting patterns functioning over large stretches of talk (accordings to Harvey Sacks), grammar focuses on more restrictive, and more detailed patterns. This bears out the relationship of grammar with very particular cultural practices, such as devising metric patterns, textual documents -- practices of language very much embedded in a local culture and devoted to either memory or repetition.