The more I'm reading about syntax, the more I'm starting to doubt my ability to judge the grammaticality of English sentences correctly... so I need some help with that. I'm a non-native speaker of English and I suspect I'm influenced by my mother-tongue (German) when assessing the grammaticality of English sentences.

Here's two complex WH-questions:

1) Where did he say he hid the treasure?

2) *Where might she ask he has been?

I'm fairly certain the first sentence is correct. If the first sentence is indeed correct, however, I fail to understand why the second sentence would be considered ill-formed (as indicated in the textbook I got the second example from; Radford: 2009). To me, their structure appears to be same.

For those acquainted with syntactic analysis:

As far as I can tell, example 1) has the underlying structure of:

[CP Where [C did [TP he [VP say [CP ̶w̶h̶e̶r̶e̶ [C ∅ [TP he [VP hid the treasure ̶w̶h̶e̶r̶e̶ ]

The wh-question marker originates in the lower VP as a complement to the verb, moves up to the lower CP boundary and then leaves the lower clause for the specifier position of the higher CP. (Movement is indicated by strikethrough. I purposely left out the other types of movement (S-Aux, VPISH) going on.)

Whereas example 2) would have the underlying structure of:

[CP Where [C might [TP she [VP ask [CP ̶w̶h̶e̶r̶e̶ [C ∅ [TP he [T has [VP been ̶w̶h̶e̶r̶e̶ ]

Again, the WH-question marker originates in the lower VP as a complement to the verb and first moves up to the lower CP boundary. It then leaves the lower clause for the specifier position of the higher CP. Radford posits that the WH-marker cannot move to the higher CP because the covert interrogative complementiser ∅ "freezes" it in place, thus producing only:

3) She might ask where he has been.

I can't see how this is the case here but not in other cases such as 1). Even if for whatever reason 1) turns out to be ungrammatical, other constructions such as 4) definitely allow long-distance WH-movement (also taken from Radford).

4) What might he think she is hiding?

[CP What [C might [TP he [VP think [CP ̶w̶h̶a̶t̶ [C ∅ [TP she [T is [VP hiding ̶w̶h̶a̶t̶ ]

Am I wrong assuming that 1) is grammatical at all? What exactly is wrong with 2)?

Any help is appreciated!


Andrew Radford, 2009: An Introduction to English Sentence Structure. Chapter 5.7, pages 174-182.

  • 1
    You might just about be able to syntactically justify Where might she ask if he has been? (or perhaps ...ask him if he has been to?), but it would be stylistically pretty awful. I'm assuming some pretty weird context where the questioner wants to know which places she might mention when asking someone else the question Have you been to X?, otherwise I've no idea why anyone would want to ask such a question. – FumbleFingers Feb 26 '20 at 16:42
  • @Lambie No, it is actually supposed to be a question. I'm not concerned with reported speech here. – esche Feb 27 '20 at 9:41
  • @FumbleFingersReinstateMonica I agree, it is an odd way to ask that question. I came up with what I consider to be a better example below with more context (which is pretty much the scenario you described). The "to"-bit is intriguing but I think it's just an prepositional addition to the copular "be" in the subordinate clause, which could've just as well been there in the original: "Where might she ask he has been to?" Doesn't really increase the acceptability of the sentence, does it? – esche Feb 27 '20 at 12:57

I think you may have "missed the forest for the trees" (den Wald vor lauter Bäumen nicht gesehen). The problem with the second sentence is the structure of things that can be asked.

The first example is indeed grammatical, and we can easily compose a corresponding answer statement:

  • (Question) Where did he say (that) he hid the treasure?
  • (Statement) He said (that) he hid the treasure there.

When we try to find the corresponding answer statement for your second example, the problem emerges:

  • (Question) Where might she ask he has been? (Ungrammatical)
  • (Statement) She might ask he has been there. (Ungrammatical)

One asks a question or predicate, such as "whether (or if) he has been there". This is similar to your native German: man fragt jemanden, ob... If we start with this form, we can find the corresponding grammatical WH-question:

  • (Statement) She might ask if/whether he has been there.
  • (Question) Where might she ask if/whether he has been?

This is grammatical, but seems like a very awkward question, so much so that it may confuse the reader/listener. We could also have the following, which is similar in appearance but grammatically and semantically quite different - I wonder if this form is also part of what's confusing you?

  • Where, might she ask, has he been?

Or, much more commonly (same structure but different pronouns):

  • Where, might I ask, have you been?

Here, "might (pronoun) ask" is an interjection, or parenthetic clause.

  • 1
    +1 for your final example. It looks a bit strange with she, but it's easy to imagine the parent of a teenage child quizzing them when they return at some ungodly hour: Where, might I ask, have you been? – FumbleFingers Feb 26 '20 at 16:50
  • @FumbleFingers I agree and added another example with a link to more information about that construction. Thanks. – TypeIA Feb 26 '20 at 16:55
  • @TypeIA Thank you, that's a useful perspective. The example with "if/whether" does sound odd to me but I see why it would be considered grammatical. What I find interesting about that construction, however, is the usage of "if/whether" in a setting where it is used to formulate an open question, and not as usual a closed one (with a yes/no answer). Consider a slighty altered version of the sentence with more context: "If she had had the opportunity to ask him, where might she have asked he had been?" Focus on the italicsed part only - it works without "if/whether"-support, does it not? – esche Feb 27 '20 at 12:24
  • 1
    @Lambie In standard varieties of English, subordinate clauses generally do not allow for inversion: "He told me where he will hide the treasure." as opposed to "He told me where will he hide the treasure.". – esche Feb 27 '20 at 12:26
  • @Lambie Again, this isn't an issue of reported speech. "Where", might she ask, "has he been?" is not a complex question because the interjection "might she ask" merely interrupts the simple WH-question "Where has he been?". You are right that somewhere in the structure of a question inversion is required to indicate interrogative mood (echoic questions aside). Still, "Where might she ask he has been?" does feature the same type of inversion as "Where did he say he hid the treasure?" (might/did + she/he + ask/say). Yet, the former yields an ungrammatical result. – esche Feb 29 '20 at 9:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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