"Ish" is a recently derived word (free root) conveying a sense of "so-so" or approximation. It is most commonly used as an adjective but occasionally as an adverb. I would not be at all surprised to see it appear in standard dictionaries within the next decade. Is there a term for the morphological process that derived the term "ish" from the derivational suffix "-ish"? Or is this an example of slang/colloquial usage?
Ish is listed in the OED as colloquial, first used 1986, and defined:
Qualifying a previous statement or description, esp. as a conversational rejoinder: almost, in a way, partially, vaguely.
(slang) used to express reservation or qualified assent ⇒ Things are looking up. Ish
The process may be called degrammaticalization.
There is a post on the Omniglot blog that uses this exact example of -ish > ish:
Degrammaticalization, a word I stumbled across on this blog today, is the process through which grammatical affixes become independent words.
A good example is ish, which started off as a suffix on words like longish, shortish, etc. Then became an enclitic – an affix that can be detached from the words it would normally be attached to, and stuck on to other words – and finally started to be used on its own. More examples of degrammaticalization include esque, ism, pro, con, anti, ette.
In Esperanto, quite a few affixes can be used as independent words. The suffix -ig, for example, indicates the cause or bringing about of action or state, e.g. blankigi, to whiten, from blanka, white. When used on its own as the verb igi, it means ‘to cause’. This appears to be a kind of deliberate, planned degrammaticalization.
The linked blog is "Mr. Verb", and the relevant post is "Degrammaticalization as liberation".
As the Omniglot post hints, this sort of thing occurs in many other languages as well.
The analysis of "degrammaticalization" is apparently somewhat debatable in linguistics: everyone agrees that specific examples like this exist, but not everyone agrees about how to explain them (or whether it is productive to try to explain them). I found a paper "Degrammaticalization: three common controversies," by Muriel Norde, (2010) that provides some discussion of this.
The term "degrammaticalization" apparently originated as a name for something that was hypothesized not to exist! Usually, we see morphemes become more grammaticalized over time, not less, and this observation caused some linguists (Norde cites Lehmann, who mentions Givón, Langacker and Vincent) to advance the hypothesis that grammaticalization is unidirectional (2).
Of course, people challenged this hypothesis with apparent counterexamples like the one in question; these counterexamples were sometimes dismissed as unimportant outliers or unexplainable freak occurences, but some linguists such as Norde have tried to treat degrammaticalization as a phenomenon worthy of study (3). Norde's paper provides some definitions and discussion of controversial points.
She discusses ish in section 6.3, "Debonding".