I recently asked a question concerning the scope of negation. I received helpful feedback from a number of linguists who frequent this forum. My efforts to discern the scope of negation continue, and in this regard, I have a new question that is related to my earlier one. However, this new question focuses only on the readings that are and are not available. I am concerned with three sentences. The first two of the three are the following:

 A. Arthur does not discipline his children because he loves them.

 B. Because he loves them, Arthur does not discipline his children.

Based upon the responses I received, it is safe to say that sentence A is ambiguous, but that sentence B lacks the ambiguity. The pertinent readings are paraphrased as follows:

 Reading 1: 'The reason Arthur does not discipline his children is that he loves them.'

 Reading 2: 'Arthur disciplines his children, but not because he loves them (but 

Sentence A allows both of these readings, whereas sentence B allows only Reading 1. This much seems to be agreed upon by everyone who responded to my earlier question.

Based upon the data just presented, I hypothesized that the negation cannot scope backwards over an adjunct. I challenged the forum to provide a counterexample. Well, a colleague of mine has in fact produced a counterexample that (in my view) clearly contradicts my hypothesis. This counterexample is as follows:

 C. ...but discipline his children because he loves them, Arthur does not.

For me and my colleague, this sentence allows the reading that has the negation scoping backwards over the cause (Reading 2). So based on this result, my earlier hypothesis can be discarded. However, there is an aspect of example C that is not clear for us: Is it ambiguous? If it is ambiguous, which readings are available? Is it ambiguous in the same way as sentence A?

My colleague and I agree that sentence C is ambiguous, but we disagree concerning the available readings. One of us believes that sentence C is ambiguous in the same way as sentence A (Reading 1 and Reading 2). The other believes that the ambiguity in sentence C is different than for sentence A (Reading 2 and Reading 3). The putative third reading is paraphrased roughly as follows:

 Reading 3: 'Arthur does not discipline his children, though he does love them.'

Thus to restate my questions as clearly as possible:

  1. Is example C ambiguous?

  2. If it is ambiguous, which readings are available (Reading 1, Reading 2, Reading 3, and/or perhaps some other reading)?

Now, a couple of days later after posting the question above, a new piece of evidence has been brought to my attention. This new evidence answers my question, I think, so I want to share it here. Observe:

 D. ...but discipline his children because he loves them, Arthur does not because
    he's never at home. 

 E. *Arthur does not discipline his children because he loves them because he's never
     at home. 

If the acceptability judgments here are accurate, they demonstrate that Reading 1 is not available for sentence C. Reading 1 interprets the subordinate clause because he loves them as giving the cause of the the absence of punishment. In sentence D, however, the cause of the absence of punishment is the second because-clause, i.e. because he's never at home. Thus Reading 1 is incompatible with the addition of the second because-clause. Reading 3, in contrast, is quite compatible with the addition of the second because-clause. Conclusion: Sentence C is ambiguous, allowing Reading 2 and Reading 3, but it does not allow Reading 1. Here are further examples that illustrate the point:

 F. ...but go to bed because I was tired, I did not because I had to finish my paper.

 G. ...but open the window because she was hot, she did not because it was hotter

 H. ...but pass the test because he had studied, Sam did not because the test was 
    simply too difficult.

These sentences are all incompatible with a reading along the lines of Reading 1. I'd appreciate any further comments concerning these additional data.

  • 1
    Example C is only a part of a sentence; what is the rest of it, and what it is intended to mean, in what context?
    – jlawler
    Sep 14, 2014 at 16:19
  • The rest of the sentence is whatever makes sense. Fronting a nonfinite VP like that is unusual; it needs to be motivated by preceding material. The question remains. Given the freedom to add whatever you want in front of the "but", which of the three readings are possible? Sep 14, 2014 at 23:28
  • Example C is an extremely marked construction. If you wanted a similar construction without ambiguity it would be "But discipline his children, Arthur does not, because he loves them."
    – curiousdannii
    Sep 15, 2014 at 1:55
  • While I prefer Reading 1 for the sentence you give, I think Reading 2 can also work: "But discipline his children, Arthur does not because he loves them, but because...". Sep 15, 2014 at 2:02
  • 1
    Your example E. reminds of one of those adventures in misnegation on the Language Log but I don't think it necessarily excludes the dual reading. The fundamental problem is that relatively very few people will be able to parse these examples properly out of context and without priming. See Ewa Dabrowska's great research - e.g. “Different speakers, different grammars: Individual differences in native language attainment”. What is the 'English' for which you are trying to discover the rules? This is where the traditional rules/structure approach breaks and needs an alternative. Sep 18, 2014 at 8:58

1 Answer 1


It seems to me that with sufficient framing, all three readings are possible (if not perhaps always straightforward):

Arthur explains things to his children kindly because he loves them. But discipline his children because he loves them, Arthur does not.

You should discipline your children out of love. Arthur beats his children out of frustration. But discipline his children because he loves them, Arthur does not.

In this culture, you must discipline your children if you love them. Arthur coddles his children because he loves them. But discipline his children because he loves them, Arthur does not.

But as fun as it is to ponder these things, I think it underscores a problem with the kind of linguistics that generates hypotheses that can be decided by acceptability of different sentences or their readings.

The Chomskean thesis that simply the views of competent native speakers are enough to adjudicate any sentence is problematic. For one, it makes the whole edifice of linguistics rest on a good night's sleep or a chance overhearing of a conversation. But more importantly it generates law-like hypotheses that impose order - similarly to physics - making language seem like an impersonal force (UG) within which we operate. But linguistics (if it must be like a science, at all) should be more like biology. Starting with a description of what is. It will establish some useful classifications but will ultimately focus on observation and documentation. What principles, it will find, it will formulate to assist itself and not the dictates of some external discipline (such as logic or computer programming).

What do we learn about language (as it is used by people) by either rejecting or supporting your hypothesis? I would much rather start with a question that asks: "What are the ways in which people negate things and what ambiguities arise in the process? How do different people deal with such ambiguities in different contexts? Are different people attuned differently to such ambiguities at different times? etc."

Examples like these might come up given an exhaustive survey but they would not necessarily deal with these examples from the perspective of scope or directionality of negation but rather from the perspective of meaning construction and context. Native speaker intuitions (and introspection) would necessarily have to play a role but they would not be the primary driving force behind the creation of knowledge.

  • Thank you for your response. I am most interested in the readings you are assigning to your examples. To be absolutely clear, Which reading goes to which example exactly? I assume Reading 1 to your first example, Reading 2 to your second example, and Reading 3 to your third example. Right? Sep 14, 2014 at 8:35
  • I agree with the criticism of Chomskyan syntax, but mostly for a different reason -- because it doesn't work. It doesn't make correct predictions. The theoretical apparatus is too complex and therefore it is not really so helpful when the goal is to better understand phenomena of syntax. I disagree with your answer, however, insofar as it seems to want to put all of syntax into the sphere of pragmatics. There are concrete and very knowable rules of syntax, e.g. many head-dependent orderings. Sep 14, 2014 at 9:13
  • @TimOsborne 1. Yes, that was my ordering of the readings. Sep 14, 2014 at 9:30
  • 2. I don't want to pull syntax completely into the realm of pragmatics but to rather blur the lines between them. The 'rules' start looking very questionable when you add meaning and context. Much better to study constructions which are much more fluid and context/speaker sensitive than traditional rules of syntax. Because 'rules' can be phrased in so many radically different ways and yet always leave things unaccounted for, I prefer to leave them behind as a way of understanding language. They are still useful for shorthand generalizations (and some pedagogy) but not with examples like these. Sep 14, 2014 at 9:39
  • Thanks again for your response and the clarification. I agree with the readings you assign to two of your examples, but I disagree with one of them. Before revealing my hand, however, I'd like to wait and see how others might respond to the question. Sep 14, 2014 at 9:42

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.