I'm inclined to say that English has two "numbers" -- not the traditional "singular" and "plural" but "singular" and "non-singular". Anything that does not represent one single thing takes the non-singular.
0.4 is not one thing (just a part of a thing), so it takes a noun marked non-singular (if the noun is a count noun). "no" is also not one thing (but none at all),
I remember people were talking on 'Ask Dr Math' about how to 'count' 0.6 apple(s) and 1.7 apple(s). It was five years ago and so my memory has got quite vague, but as far as I can remember, what they were saying was that English people say 0.6 apple and 1.7 apples, while French people say 0.6 apple and 1.7 apple.
I made a Google survey for the choice between X and Xs. For the values less than one, it appears to depend on personal preference. 0.1 mile/miles per .. : 223/148 0.1 hour/hours and .. : 98/109 0.5 mile/miles per .. : 136/264 0.5 hour/hours and .. : 468/635 1.1 mile/miles per .. : 13/161 1.1 hour/hours and .. : 16/195 1.7 mile/miles p
I did a new survey and the results were : 0.7 mile/miles per...:10/55 and 0.7 hour/hourd ..:17/99. This sounds to indicate the proximity actually works on the choice. But the proximity's influence seems not so strong in the case 0.1 and 1.1. Is it because you say rather 'one tenth' than 'point
It seems to me as if the feeling we have against saying 'miles/hours' after '1' (e.g. 223/148, 98/109 for 0.1) starts to disappear, once the figure itself is obviously greater than 1.