In traditional linguistics literatures there is a clear separation between words and non-words. Words are basically what you'd find in a dictionary. But in todays world you find all kinds of word-like "things" in tweets and other social media, such as you could imagine:

I'm so excited, yay!!ah!!woooohoo:):):Dboom
I've been w#ndreeeng what this means.
I don;t knw if this <asdfasdfasdf> sense.
Maybe the wordsarestucktogeth ...er.


  • yay!!ah!!woooohoo:):):Dboom is basically enthusiasm
  • w#ndreeeng is "wondering" with a sort of sounded out part (reeeng) and the # for playfulness.
  • don;t is misspelled
  • knw leaves out some letters
  • <asdfasdfasdf> can be interpreted as "makes" for "makes sense".
  • Words are stuck together
  • But also split apart.

These are just the top of my head. But I am wondering how to structure them linguistically. Such as, calling them just "words", but that doesn't make sense. Calling them "things" is too general. Maybe "clauses", but that is usually a set of words. So wondering if there are any formalisms around this sort of stuff. Wondering how to treat them.

You can also have more normalized "non-word" structures such as:

This is a #hashtag.
This is a https://linguistics.stackexchange.com link.
This is a @username.
This is an example.domain.com.

Wondering what those types of things are called when they are found in text/writing/language.

  • "Such as, calling them just "words", but that doesn't make sense." Why? Words are just things you put spaces around ;)
    – curiousdannii
    Jul 30, 2018 at 4:12
  • 3
    In linguistics there's a clear separation between words and non-words? Which textbook are you reading? :p Jul 30, 2018 at 4:18
  • I am talking about e.g. punctuation, paragraphs, phrases, as opposed to words.
    – Lance
    Jul 30, 2018 at 4:22
  • 1
    In terms of parsing you could just call them all tokens.
    – curiousdannii
    Jul 30, 2018 at 11:51
  • "Words are just things you put spaces around" That's just one definition of a word and the worst one. That's the concept of a graphematic word which is useful when you want to count "words" of a text or are interested in spelling conventions. Dec 15, 2018 at 15:58

1 Answer 1


Those things are considered special parts-of-speech in corpus linguistics. There are several lists for parts-of-speech in general use, and in Universal Dependencies they are called "Symbols". Some of them may also be classified as Interjections, Particles, Other, or something else.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.