"Here he comes", "Here comes he" : The order of pronoun and verb in inversion

Information packaging constraints

Here comes he.

This sentence would usually be considered malformed, whereas Here comes Bob would be considered perfectly correct. Both are cases of what is known as Subject-dependent inversion.

Subject-dependent inversion has to do with the linguistic concept of information packaging. In other words it is about how we arrange the order of words and phrases in a sentence.

The vast majority of non-canonical phrase orderings in English are due to the status of particular words and concepts in terms of the current discourse; in particular, whether items are considered as old or new in the conversation. Most importantly they make use of the fact that we prefer to put things that have already been discussed in the recent discourse at the beginning of the sentence. New concepts and information tend to come at the end. It is the end of the sentence which carries the most emphasis, not the beginning.

A general constraint on the felicity of most types of phrase reordering, therefore, is that phrases that are preposed should if possible be discourse old (already have been talked about) or be in the speaker's or listener's immediate vicinity. The general motivation for doing this is so that some other phrase will then appear towards or at the end of the sentence where it will receive the most emphasis, and be easier for the listener to process. For this reason there is an even stronger constraint on elements which are propelled from their canonical position towards the end of the sentence: these items should usually be discourse new. They should not already have been being talked about. There are, of course, some exceptions to this 'rule'. For example, items that the speaker wants to give contrastive stress to may be moved to sentence final position, where they will receive the most emphasis.

Locative and temporal adjuncts (read adverbials) buck this general trend, in that whilst putting new information at the beginning of a sentence seems to put a lot of strain on the listener, the preposing of locative complements seems to make it easier to process sentences where nearly all the information is new.

These types of considerations are usually what motivates the use of (long) passives, existential constructions, so-called it clefts and, of course, subject-dependent inversion. Compare the following pairs, where the first has a canonical phrase ordering and the second uses some alternative form of information packaging:

1. a. I've been studying the Mona Lisa. Leonardo Da Vinci painted the Mona Lisa.

   b. I've been studying the Mona Lisa. The Mona Lisa was painted by Leonardo Da Vinci.

The (b) example here is better. The first is rather stilted. The reason the passive sentence works well is that the Mona Lisa links back to the first sentence, whilst Da Vinci, which is new to the discourse, takes the prominent position at the end.

2. a. A cafe is on the corner.

   b. There's a cafe on the corner.

   c. On the corner is a cafe.

Example (a) feels very awkward here. The reason is that a cafe is new information. We know this because of the use of the indefinite article, "a". It's awkward and difficult to process in this position. Both (b) and (c) examples are fine. The existential, there is, construction gives the reader a split second to anticipate the introduction of a new entity into the discourse. The (c) example uses subject-dependent inversion. This places the locative adjunct, which is relatively easy to process, at the beginning of the sentence. The phrase a cafe, in contrast, gets transported to the end of the sentence where it receives the most emphasis (compare the prominence of cafe in this example, with its relatively weaker status in the existential).

Here comes he.

The reason that this sentence will strike most readers as odd is that he occurs in a very prominent position at the end of the sentence. This position would normally be reserved for new entities, especially if we have specifically changed the normal canonical word order of the sentence. The fact that we are using a pronoun here means that we must already have been talking about this man in the previous discourse. Having a discourse old word shunted into this position goes against our natural language expectations and seems to jar. However, we might possibly be able to manufacture a scenario where this might be acceptable. For example imagine the following exchange:

• The only person who could save us now is Bob.
• Oh, look! Here comes he!

In the exchange above, we might possibly consider the sentence grammatical. The reason for this is that in some sense Bob really is arriving anew in the discourse. In the original sentence Bob was a displaced entity, far from the immediate environment. Bob's entry into the immediate vicinity of the speakers makes him new and newsworthy again. In any case, whether or not this is the main reason for the increased acceptability of this inverted sentence, the fact is that Bob is centre stage in this sentence and can therefore be moved to this emphatic position.

One other detail, putting aside analysis, is that English shares a certain preference for putting the verb second in some phrases with German, where it is not just a preference but a rule in most cases. In English, we have retained that pattern in many cases where we are never taught that this is any kind of rule (because it isn't anymore).

This shows its head in phrases like:
"Long live the king!"
"Long have you hunted me. Long have I eluded you." (The Return of the King)
"Now is the winter of our discontent. Made glorious summer by this sun of York..." (Richard III)
"Many are the afflictions of the righteous: but the LORD delivereth him out of them all." (Psalm 34:19)
"Gone are the days when... "
"Fain would I dwell on form; fain, fain deny." (Romeo and Juliet)
"Never have I ever..."
"Pop! goes the weasel!"
"This have I done for my true love."

In German, all of these would have to have the finite verb in 2nd place and any non-finite verbs all at the very end of the clause. We lost that last bit long ago, but this verb-second feel is very evident in German.
It is especially pronounced when there is a modal verb in second place and other verbs later in the sentence. The fact that it sounds archaic would suggest that it comes to English directly from German.

Regarding pronouns, English shares another detail with German: phrases like "Pick it up." In all of these kinds of phrasal verbs in English, you can put the object either between the verb and the preposition/adverb or at the end; but a pronoun must be in the middle, never at the end:
"Pick the banana up." = "Pick up the banana."
"Pick it up." =/= *"Pick up it."