Here's a paragraph from this EFL Magazine article "SUBJECT RAISING: DO YOU HAPPEN TO KNOW?" (2015):

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.

(Emphasis mine.)

This paragraph seems to suggest that 'raising' is an outdated concept in modern linguistics (although the author still finds it useful).

Is this true?

If so, how does 'modern linguistics' explain what is called 'raising' constructions?


For example, doesn't HPSG do away with the concept of 'raising' altogether? I don't know if HPSG is one of the modernest linguistics theories, but it seems fairly new.


This paper on HPSG "Lingering Challenges to the Raising to Object and Object Control Constructions" has this description of HPSG regarding example (45):

The HPSG account shares with the overt raising account the assumption that ‘Marcia’ in (1), repeated here as (45), appears in the main clause in the surface string.

(45) Cindy believes Marcia to be a genius.

It differs from the movement accounts, though, by assuming a monostratal syntax, which means that though ‘Marcia’ is the object of ‘believes’ in the phrasal syntax, it is associated with the syntactic and semantic features of the embedded predicate (‘to be a genius’) by a kind of complete phrase coindexing called structure-sharing, and not movement. In a sense, the NP ‘Marcia’ is equally associated with the main clause verb and the embedded clause verb; but in the surface string it is in the main clause.

This sounds like HPSG doesn't posit movement (or raising for that matter) to analyze (45).

  • The concept itself has not disappeared - it's just what triggers this phenomenon is explained differently now (feature checking).
    – Alex B.
    Mar 21, 2019 at 2:43
  • @AlexB. Isn't the concept of 'raising' based on the distinction between deep structure and surface structure? Or are you saying that the concept of 'raising' is compatible with the elimination of the distinction between deep structure and surface structure?
    – JK2
    Mar 21, 2019 at 3:06
  • 1
    Observed language data is more or less the same, it's how you analyze your data that changes with time. In generative syntax (the MP), movement still exists.
    – Alex B.
    Mar 22, 2019 at 22:36
  • 1
    On the matter of why there are doubts about transformations, I agree with @jlawler, and my own opinion that there are no transformations has only to do with grammar and has nothing to do with any putative evidence about what does or doesn't happen in people's brains.
    – Greg Lee
    Mar 23, 2019 at 23:01
  • 1
    The standard test for synonymy of two sentences is logical: two sentences are synonymous if, whenever one is true, then the other one is true and whenever one is false, then the other one is false. Logic only has two "meanings": True and False. Presuppositions, entailments, implicatures, and other fine-tuned context-dependent interpretations count as pragmatics.
    – jlawler
    Mar 26, 2019 at 14:32

1 Answer 1


No, Raising is alive and well, but the conception of Raising as a transformation is moribund, because transformations are no longer accepted. So, if we believe in Raising, and Raising is a transformation, where does that leave us? In a word, confused.

Confused, but at least open to the possibility that Raising exists but is not a transformation. Well, transformational grammar is not the only theory under the sun, after all, so we should look to some other theoretical account of Raising. My favorite candidate is Relational Grammar, one of whose proponents, Paul Postal, it happens, wrote the book on Raising.

HPSG, which you mention, is a revision to GPSG, and GPSG in a certain (rather trivial) sense contains Relational Grammar, since it attributes a "GR" (Grammatical Relation) to every NP, and offers a description of Raising. However, in my opinion, these extensions of CFG (Context Free Phrase Structure Grammar) are too artificial to be true.

Thinking of Raising as an upward movement in a tree structure, I myself offered a theory here, which I called "2psg", which describes Raising as a change in grammatical relation, or obliqueness. (Construing grammatical relations as degrees of obliqueness is also found in HPSG.)

  • As far as I know, HPSG abandoned movement for structural sharing. If you've abandon movement, doesn't it mean that you've abandoned Raising as well?
    – JK2
    Mar 22, 2019 at 1:39
  • 1
    No, it doesn't mean that. As was shown in GPSG, the relationship between constructions without movement and those with movement (in TG terms) can be described completely within CFG (which has no movement mechanism). It's not obvious, I know. Perhaps you could ask a question about it?
    – Greg Lee
    Mar 22, 2019 at 1:56
  • By the way, I don't know what structural sharing is.
    – Greg Lee
    Mar 22, 2019 at 2:02
  • 1
    @JK2: Welcome to linguistics, where terminological confusion is the norm, lol. Raising was originally proposed as a transformation mechanism, i.e. as a descriptive tool, and this seems to be the norm in older papers, but the term has since been co-opted by linguists to refer to the phenomenon rather than the descriptive tool. Mar 22, 2019 at 8:40
  • 1
    Nothing of substance can ever follow from definitions in an empirical science. The Raising analysis can be expressed as a relationship between tree structures of sentences. If a tree with a verb and sentence complement is grammatical, then a certain other tree, with object and infinitive, will be grammatical. The relationship between the trees can be described by moving things around in the first tree to obtain the second tree, but it needn't necessarily be. Instead, it can be described as a relationship between the PS rules that generate the first tree and those that generate the second.
    – Greg Lee
    Mar 22, 2019 at 8:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.