So here’s a common refrain:
“Very unique is an example of perversion not uncommon to every-day usage.” – The Worth of Words, Ralcy Husted Bell, 1902.
“A thing is unique (the only one of its kind) or it is not. Something may be almost unique (there are very few like it), but nothing is ‘very unique.'” – Paul Brians
If we agree with the grammarians, very unique is an abomination, for unique means one-of-a-kind, and how can something be more or less one-of-a-kind? You’ll note that this is a instance of argument #2: that logic dictates that some bit of usage is impossible. (And here I want to point out that if “very unique” is really logically rubbish, then why do so many otherwise logical people understand what it means?) However, the idea that uniqueness can be graded and different things can be more or less unique is actually logically reasonable.
For starters, everything is truly one-of-a-kind. Just as there are no two identical snowflakes, there are no two identical objects anywhere in the world (excluding, perhaps, microscopic particles). Identical twins are different from each other. So are two boxes of cereal on the shelf at the supermarket (one will certainly have more bits of cereal than the other). So if everything is unique, why not just say it’s a meaningless word and do away with it altogether? Because, within our minds, some things are sufficiently similar to be lumped together in our minds as identical for certain purposes. This leads to an argument that X is unique if and only if there is no other thing of the same kind as X that can be considered identical to X.
The problem with this definition is that different things can be considered identical in different contexts. Take the case of Harry Potter. There’s two swords of Godric Gryffindor. One is made by goblins and is therefore imbued with special powers, and the other is a clever replica that looks almost the same, cuts the same, but lacks the magical powers. To the uninitiated (e.g., Bellatrix Lestrange), the two swords look and behave identically, but to a wise goblin (e.g., Griphook), they are noticeably different. Thus, to Bellatrix, the swords are not unique, but to Griphook they are. I am amazed I managed to make a grammatical point using Harry Potter. So if a single item can be viewed as unique or not by different proportions of the population, does it not make sense to say that things that are unique to more or less of the population can be referred to as more or less unique?
Secondly, something can be one-of-a-kind, with the “kind” having a variety of different sizes. To say that a person is unique is to say that he or she is one amongst six billion. To say that a mole on my skin is unique is to say it is one amongst a hundred (the number of moles I have). To say that a particular insect is unique is to say it is one amongst 300 quintillion (an estimated number of insects in the world). So it sure seems that intuitively a “unique” insect is more unique than a “unique” mole. So that insect, in my opinion, is very unique.
In a similar vein, one unique thing can be different from similar objects in more or fewer ways. In math-speak, it could be unique on different numbers of dimensions. A sculpture, for instance, could be unique in size, shape, material, placement, inspiration, sculptor, price, location, color, time-period, or intention. Suppose sculpture X is unique as the only tin sculpture of the Renaissance, but sculpture Y is unique only as the only black marble sculpture weighing more than a thousand pounds done by Georg Cantor in 1897 to protest Prussian treatment of the French. Sculpture X is then one of more kinds than sculpture Y is one of, and furthermore, sculpture X is one of a kind in a much more general kind than sculpture Y. So it again is logically possible to argue that sculpture X is more unique than sculpture Y.
My point is that very unique is not logically reasonable only to someone who interprets the definition of unique in a downright contrarian way. There are many possible ways for two things to be more or less one-of-a-kind. So it cannot be claimed that logic precludes the use of very unique.
Furthermore, this whole argument is precarious anyway, as it assumes that the sole definition of unique is “one-of-a-kind”. The Oxford English Dictionary does give this as a definition for unique, but it also notes a second definition that has been in use since the 18th century and has been “very common” since the middle of the 19th century: “standing alone in comparison with others, freq. by reason of superior excellence; unequalled, unparalleled, unrivalled”. This is attested in writing as early as 1809, by R. K. Porter: “As it was thoroughly unique…”, and further examples from the 19th century occur in force in a full text search of Google Books for “very unique”. This undermines argument #1: even authors of the good old days succumbed to the reasonable temptation to modify unique.
Summary: Unique can be modified without complaint; better writers than me have done so for almost 200 years and there is no reasonable logical argument against it. However, unique is, in my opinion, overused, in part because everything is unique in some sense. As such, I would recommend avoiding unique (modified or not) when you can say it another way.