When we build sentences, we don't always set our words down into a syntactic structure in the exact order they show up in when we say them. Instead, we put words down in the positions in the sentence where they'd be interpreted normally: say, subject followed by verb followed by object in English. Then, we move around our words to make the kinds of sentences we want, like questions, according to the rules of the language you're speaking.
How can we know that this is happening? We see effects on where words have moved from that makes it seem as if some small, invisible thing is left behind in the site where the moved word originally was. We call these silent remains traces. Traces prevent words from filling in the space where the word was originally located with other words, and explain why some of the time, we can't make abbreviations by contracting two words that we'd normally be fine with shortening, like changing want to to wanna.
Also, these movements allow us to account for some of the variation we see in the world's languages. We want to say that underlyingly, languages all rely on the same syntactic architecture, and by saying that some languages move kinds of words that other languages leave in place, we can explain a lot. So for example, French and English word order differs for where negation and adverbs go, because in French, the verb moves up out of its original position, whereas in English, it stays in place.
In our video, we discussed some of the real-world effects that moving words around in a sentence can have. If we turned the sentence in (1) into the question in (2), we would expect to be able to pronounce “want to” as “wanna”. And we can do just that, giving us the question in (3). That's because there’s nothing to stand in the way of “want” and “to” cozying up to one another — no words and no traces.
(1) I want to defeat the ghost.
(2) Who do I want to defeat?
(3) Who do I wanna defeat?
Things change when something falls between them, even if we move that something out of the way. So, when we turn the sentence in (4) into the question in (5), we can’t pronounce it as it’s written in (6). Note that we use an asterisk to show that a sentence is unacceptable.
(4) Annie wants Parley to defeat the ghost
(5) Who does Annie want to defeat the ghost?
(6) *Who does Annie wanna defeat the ghost?
This is a pretty good piece of evidence that there’s something standing between “want” and “to”, even if it can’t be seen with the naked eye. But you might be wondering if there’s even more concrete evidence that shows there really is something left behind when a word moves from one part of a sentence to another. And there is: we just need to look at kids again!
An alternative to the idea that traces are left behind when words move out of their starting positions is the idea that words don’t move at all, but instead make copies of themselves and send those along in their place. This is called the copy theory of movement. Then, when we’re done building our tree, we only pronounce the last reproduction of a word, the one that made the last movement. That leaves all the other copies that came before it, including the original one, unspoken.
So why might we think this idea could be more accurate than just having traces? Why do we think that these unspoken copies, rather than traces, are the things sticking in between words like “want” and “to” in the sentence in (6) above? Well, sometimes these extra copies get pronounced out loud — like a trace you can actually hear. In child speech in English, for example, both a copy and its original can find their way into a question.
(7) Who do you think who is in the box?
This might not be the kind of thing an adult would say, after they learn that they aren’t supposed to be pronouncing every instance of the same word in a single sentence. But the fact that children do it gives us even more reason to believe the idea that questions don’t simply appear out of thin air, but are actually very closely related to their non-question counterparts. Kids make the extra copies and do the movements right, but then they forget to not say the later versions. And that gives us more than one version of a question word like who in a given sentence.
When we think about movement this way, as having full, unspoken copies, we can understand not only why the sentence in (6) is bad — after all, how could “want” and “to” ever become “wanna” if there’s a “who” wedged between them — but why sentences like the one in (7) exist at all.
So how about it? What do you all think? Let us know below, and we’ll be happy to talk with you about what moves around in what different languages, and the traces or copies that get left behind. There’s a lot of interesting stuff to say, and we want to hear what interests you!
Previous Topic: Next Topic:
Linguistic Theory of Relativity Coming soon!