Can we improve spaCy patterns; allow to match on single or multi-token?

[sam-writer]

So often I want to match word1 word2 or phrase_1 phrase_2 phrase_2 and suggest replacing them with replacement. Because the spaCy matcher is token-based, this requires multiple match rules... that seems silly. e.g. "replace dumb, mute, or non-verbal with unable to speak.

Can we have a shorthand for this instead of having to have as many matches as different token counts?