4 ms·

The way the study is presented makes the reasoning sound somewhat circular: Language A is optimal because the dependencies are short. The dependencies are shor

The way the study is presented makes the reasoning sound somewhat circular: Language A is optimal because the dependencies are short. The dependencies are short because that's optimal. Not sure what we learn from that. I also find statements like "language A is more optimized than language B" questionable because (i) it's not clear how to measure optimality and (ii) there are no good reasons to assume that some languages evolve to be flawed, the pressure to perform well is just too high. Anyway, I suppose that these are just problems of the dumbed down press release. Looking forward to see the actual paper.

jxramos 11y ago
I think the criterion for optimization came about from this line... “If there is a large amount of time between one word and another related word, that means you have to hold one of those words in memory, and that can be hard to do.” They're basically saying you can pop things off your mental stack more readily with smaller DLM languages.
- tmalsburg2 11y ago
  Just had a look at the paper. They indeed assume that long dependencies are hard and that languages therefore tend to optimize for short dependencies. They also argue that some languages (like German) may have longer dependencies because other aspects of their grammar make it easier to process those dependencies than would be the case in short-dependency languages like English. This suggests that languages that are more optimized for dependency length may be less optimized with respect to some other aspect of processing. Languages may thus make different tradeoffs with respect to the various sources of processing difficulty. This means that German may have seemed hard to Mark Twain not because it is intrinsically harder but simply because it makes different tradeoffs than English.

4 ms·

The way the study is presented makes the reasoning sound somewhat circular: Language A is optimal because the dependencies are short. The dependencies are shor

by tmalsburg2 11y ago

jxramos 11y ago
I think the criterion for optimization came about from this line... “If there is a large amount of time between one word and another related word, that means you have to hold one of those words in memory, and that can be hard to do.” They're basically saying you can pop things off your mental stack more readily with smaller DLM languages.
- tmalsburg2 11y ago
  Just had a look at the paper. They indeed assume that long dependencies are hard and that languages therefore tend to optimize for short dependencies. They also argue that some languages (like German) may have longer dependencies because other aspects of their grammar make it easier to process those dependencies than would be the case in short-dependency languages like English. This suggests that languages that are more optimized for dependency length may be less optimized with respect to some other aspect of processing. Languages may thus make different tradeoffs with respect to the various sources of processing difficulty. This means that German may have seemed hard to Mark Twain not because it is intrinsically harder but simply because it makes different tradeoffs than English.