Another first (as far as I can tell) for Bain was the idea that prepositions do not suddenly turn into "conjunctions" when they introduce a clause. On p. 64 of An English Grammar (1863), Bain mentions "prepositions governing clauses". This is the seed of an idea that is taken up by Otto Jesperson in the first quarter of the last century and most recently in The Cambridge Grammar of the English Language (CGEL), which argues against the traditional view that has prepositions hopping from class to class as in the following three sentences:
- I stood before the door. (preposition)
- I stood before I spoke. (conjunction)
- I have stood here before. (adverb)
Similarly, nouns can stand alone or license a wide range of internal complements; yet we still allow that they are nouns. There is nothing especially confusing about this. The same goes for adjectives. Indeed, it seems far less confusing to be told that prepositions are defined mostly by where they come in the sentence and the type of information they convey, as we do with most other word classes, than to be told that with prepositions the defining property is mainly the constituent that it licenses.
The CGEL continues the argument by pointing out that if it could be shown that there were some other properties that changed when one of these words takes a different complement (or none at all), that would argue for the traditional view, yet this is not the case. For example, the same modifiers are allowed:
I had begun to walk one month before my first birthday.
I had begun to walk one month before I crawled.
I had begun to walk one month before.
I moved straight inside the house.
I moved straight inside.
If it were only a small number of words that straddled the line between prepositions and adverbs, this too would argue against this position. But the overlap is quite considerable and includes everything from core members such as 'in' and 'on' to oddballs like 'apropos' and 'notwithstanding'. Where the majority of the members of a group are falling into two (or three) categories, it seems reasonable to reassess the criteria by which they were assigned.
Even if we do allow for this needlessly strict definition of preposition, why then are we so promiscuous with our characterisation of adverbs? According to the CGEL, while there are instances in which adverbs do function as complements(e.g., "They treated us KINDLY."), generally speaking adverbs functions as adjuncts. There are no ~ly adverbs that can function as a goal complement to replace the prepositions in the following sentences:
They went ABOARD.
Please bring them DOWNSTAIRS.
She came BACK last night.
I can't get it INDOORS.
Nor does 'be' license ~ly adverb complements, though it does license prepositions. Consider, then, the following:
They are aboard.
They are downstairs.
If we accept the traditional view that these are adverbs, what does that say for the traditional characterisation of be complements?
It seems, then, that adverb is simply a label that we give to words that don't fit anywhere else. This is not good scholarship. Pedagogically, the result of accepting a broader definition of prepositions, following Bain, Jesperson, and the CGEL, would be a much simpler, more consistent explanation for students. They would still have to deal with learning which prepositions license which type of complement--and which are intransitive--but they would not be faced with the confusing range of "exceptions".
I can't imagine other disciplines in which, all evidence to the contrary, teachers cleave to the traditional view simply because it is tradition, and Bain's 1863 work isn't exactly fresh. I think we language educators need to realise where we're wrong and accept a better analysis when it presents itself.