(function(doc, html, url) {
var widget = doc.createElement("div");
widget.innerHTML = html;
var script = doc.currentScript; // e = a.currentScript;
if (!script) {
var scripts = doc.scripts;
for (var i = 0; i < scripts.length; ++i) {
script = scripts[i];
if (script.src && script.src.indexOf(url) != -1) break;
}
}
script.parentElement.replaceChild(widget, script);
}(document, '
Automatic methods for scoring the complexity of people's language
What is it about?
To communicate, we combine words into sentences, which can have a simple syntactic structure or a more complex one. In the current publication, we explore ways for automatically quantifying syntactic complexity in speech. We studied syntax of older and younger adults. First, we showed that older people indeed tend to produce fewer complex syntactic structures. Then, we explored different automated ways for scoring the complexity of speech based on its syntax. Our investigation included seven ways that were previously used in the literature, and one new way that we developed. We showed that our new method was the most sensitive one in capturing the differences in level of syntactic complexity between older and younger adults.
Why is it important?
This should be useful for anyone who is seeking for a way to quantify syntactic complexity. There is currently no standard way for scoring texts or speech for their level of syntactic complexity. Our study is the first most comprehensive comparison of different metrics for quantifying syntactic complexity. Our findings show how consistent these metrics are with each other, and rank them based on how well they performed in distinguishing between two groups of different syntactic levels. In addition, our study also evaluates the accuracy of syntactic complexity analysis using automatic speech recognition (ASR) and discusses the required challenges for using ASR in a fully automated syntactic complexity scoring tool.