Subject Raising: Do You Happen To Know?

Subject raising is my all-time favourite English grammar structure. It’s one of the craziest things you can do in English, and yet it’s hardly ever mentioned in course books. For most teachers and learners, it’s invisible; all you see is a bunch of idiomatic expressions and weird exceptions. But once you start noticing the patterns, subject raising is everywhere.

As the name suggests, subject raising involves raising the subject from a lower part of a sentence (in other words, a subordinate clause) to become the subject of a higher part of the sentence, usually the main clause. Here’s a simple example:

  1. It seems [that you’re worried].
  2. You seem [to be worried].

The subject of sentence 1 is ‘it’, a dummy subject, in other words, a subject with no meaning. The dummy subject simply fills the role of subject in a sentence, which can’t usually be left empty in English. Another example of a dummy subject is the meaningless ‘it’ in sentences like ‘it’s raining’, ‘it’s dark’ or ‘it’s ten o’clock’. However, the that-clause in 1 has a real subject, ‘you’, which certainly has meaning.

Sentence 2 means the same as sentence 1. The only difference is that the subject of the sentence is now ‘you’. It’s almost as if the subject from the that-clause has been lifted up to become subject of the whole sentence. In the process of losing its subject, the that-clause has been reduced to a mere to-infinitive. This process is called subject raising (or, more precisely, subject-to-subject raising), and the subject of sentence 2 is called a raised subject.

Understanding the structure

Not long ago, most linguists believed there really was a set of processes in our brains called ‘transformations’, where words moved around inside sentences, to turn simple statements into things like questions, negatives, passives, etc. This idea seems to have fallen out of favour in modern linguistics, which feels like a shame to me – I like the idea of words whizzing around inside our brains. But whether or not it really happens in our brains, I think the image of subjects being raised to different parts of sentences is still a useful way of understanding the structure.

Watching DVDs on your microwave

Why would anyone want to raise a subject? Well, for a start, it cuts down on the number of words. It also eliminates the need for a meaningless ‘it’ at the beginning, which seems like an improvement to me.

But for me the best thing about raising is that it’s so crazy. In a logical world, we shouldn’t be allowed to move subjects from one part of a sentence to another! It doesn’t make any sense … and yet it works. It’s as crazy as plugging your DVD player into your microwave oven instead of your TV and expecting it to show films – except that it really does work with subject raising.

Let’s have a look at some more examples.

In each case, the (b) version is the one with the raised subject.

3a. It’s likely that they’re waiting in the office. b. They’re likely to be waiting in the office.

4a. It appears that we missed them. b. We appear to have missed them.

5a. It happened that I saw Dave yesterday. b. I happened to see Dave yesterday.

A few interesting points come out of these examples. Notice that the progressive form in 3a (they’re waiting) leads to a progressive infinitive in the raised version (to be waiting). Similarly, the past form in 4a (missed) leads to a perfect infinitive (to have missed). But in 5a, the past form (saw) doesn’t lead to a perfect infinitive – it’s just a normal infinitive (to see, not to have seen). The main verb (happened) is in the past tense, and that’s enough to make it clear that the whole sentence is about the past.

There’s sure to be trouble …

In all these examples so far, a meaningful subject has replaced a dummy subject (it). But this isn’t always the case, as the examples below show:

6a. I’m certain she’ll be here soon. b. She’s certain to be here soon.

7a. It’s likely that it’ll rain tonight. b. It’s likely to rain tonight.

8a. It is said that there are hundreds of fans. b. There are said to be hundreds of fans.

9a. I’m sure there’s a solution. b. There’s sure to be a solution.

In sentence 6, we’ve got one meaningful subject (she) replacing another meaningful subject (I). This is much less common, and is mostly restricted to sentences with ‘sure’ and ‘certain’, where the lost subject, ‘I’, is still obvious from context. In sentence 7, in contrast, we’ve got a dummy subject (the ‘it’ from ‘it’ll rain’) replacing another dummy ‘it’. The (a) and (b) versions both appear to start with the same word, ‘it’, but it’s a different ‘it’ in each case. In sentence 8, we’ve got another dummy subject, in this case the ‘there’ from ‘there is/are’, moving up to replace the dummy subject ‘it’. Finally, in sentence 9, we’ve got a dummy subject ‘there’ replacing a meaningful subject, ‘I’.

Did you happen to notice?

So when can we raise subjects in this way? Well, it’s limited to about a dozen or so structures:

  • verbs of seeming: appear and seem (e.g. You seem/appear to be right). But not look, sound, feel, etc.
  • adjectives of probability: sure, certain, likely, unlikely (e.g. You’re likely/sure to be right). But not probable, possible, uncertain, inevitable, etc.
  • passive reporting verbs: be said, be thought, be known, be believed, be rumoured, etc. (He’s said/known/thought to be dangerous).
  • passive verbs of expectation: be expected, be required, be supposed (e.g. He’s expected/supposed/required to be here soon)
  • verbs of happening: happen (e.g. If you happen to see Jim, …) and turn out (e.g. We turned out to be the only guests). But not occur.

Some of these structures (e.g. be supposed and happen) are rather idiomatic: their meaning has moved away from the meaning of the simple, unraised version:

  • I happened to see Jim” means “It happened by chance that I saw Jim”.
  • You were supposed to come” means “it was expected that you would come” – the verb suppose by itself doesn’t have the same sense of expectation.

Nevertheless, it’s still useful to think in terms of raising for these expressions, even if the meaning has drifted.

The structures with passive reporting verbs are rather cool. There’s no active verb equivalent to these raised structures – for example, we can’t say “They think him to be dangerous”. The verb be rumoured is doubly cool, because it’s one of the only verbs in English with no active form at all: we can say “it was rumoured that …” but not “they rumoured that …”.

One of my favourite words

Finally, I’d like to mention one of my favourite words, bound, as in “You’re bound to pass the exam”. This word seems to be almost untranslatable into other languages … until you realise that it’s only used in raised subject constructions. It belongs in the same family as sure, certain, likely and unlikely, except that there’s no non-raised version. We can’t say “It’s bound that you’ll pass the exam”. In other words, bound is a structure with a permanently raised subject. Nice word, huh?

So what do you think? Should we teach subject raising more prominently and at lower levels? Should we treat it as a cool rule or a bunch of idioms and exceptions?

Related Topics

Leave a Reply

Your email address will not be published. Required fields are marked *

4 Responses

  • hire a killer

    Raising-to-subject is triggered by predicates which take only one argument - a proposition - and do not assign an external (subject) theta-role (i.e., they do not semantically select their subject dependent, but instead raise one from a sentential complement with a caseless subject position), as well as by passive forms of Exceptional Case-Marking verbs (able to govern the embedded subject position of their infinitival complement, such as 'believe'). It is because of this that the subject of the matrix clause may be semantically empty. These class of verbs need to be distinguished from the so-called control predicates, which take two arguments and do not involve DP movement, e.g.: "Early in 1066 they were reluctant [PRO to accept Harold Godwinsson as their new king]." which is why clausal subject or extraposition are not possible with this latter class of predicates. (Technically speaking, under a generative analysis, the subjects in raising constructions undergo double raising: first from within Spec-VP to subject position – Spec-TP – of the infinitive complement clause in order to satisfy the Extended Projection Principle requirement, and then to Spec-TP of the main clause for the DP to get Case.) The verb 'look' *may* appear in the raising construction: "The landing looks to have been rough as the sleigh has dug a deep fifty foot long trench" (COCA) "Tod looks to have missed with both his shots." (BNC) and as attested by several hundred further examples in language corpora. Also, raising constructions with passive reporting verbs may have active equivalents; for instance, "they thought him to be" returns 8 instances in the BNC (e.g. "I had always thought him to be egotistical and attention-seeking"), 14 in COCA ("We thought him to be the original voice of dissent"), and 18 on the pages of The Guardian. Besides economy, one of the motivations for using a raising construction is maintaining thematic progression or shifting focus. As raising occurs in several languages even from families typologically or phylogenetically remote from West Germanic (e.g. West Slavic), showing analogies between the constructions in both may facilitate students' acquisition thereof. Otherwise care should be taken not to make the picture look too complicated and technical, as that may discourage the learners and raise their affective filter.

    29/04/2015

  • Caio Albernaz

    Jeremy, this is definetely a very interesting topic. I used raised subject many times, but I'd never bumped into its grammar structure and explanation. Awesome! By the way, I love using "Cambridge English for Engineering" with my ESP students. Regards, Caio

    06/05/2015

  • Jeremy Day

    Hi Michał Thanks so much for this clarification and analysis. I think it's a really interesting subject, and I can see you have a far deeper understanding of it than me. As you'll have noticed, my approach in this article (and similar articles in future) is as an enthusiastic grammar-obsessed English teacher, rather than an expert or a linguist. If I can share a bit of that enthusiasm with other non-experts, (without getting the technical stuff horribly wrong), I'll have done my job. I want to make certain grammar topics as accessible as possible without dumbing down. Anyway, thanks for correcting my sweeping statements about 'look' and 'think'. I guess I wouldn't use those structures myself, but your examples sound like fairly typical journalistic style. That's the benefit of using a corpus, I guess. I don't know much about raising in other languages, but I'll start listening out now for examples in Polish. (I live in Poland, you see.) Finally, yes, I'm sure you're right about the affective filter. In my experience as a teacher, students are generally happy to learn about things like raising when the focus is on helping them to get their heads round a strange expression they've encountered, and when the teacher presents it clearly and enthusiastically. But I'm sure a cold, dry lesson on this topic would turn a group of teenagers off immediately! Thanks again for you great comments.

    06/05/2015

  • Jeremy Day

    Hi Caio That's great to hear. I'm glad you found it useful ... and that you agree with me that it's a cool topic. Also, thanks for the kind words about Cambridge English for Engineering. I'm very proud to have been part of the team that worked on that one.

    06/05/2015