Phonological Ambiguity and Autocorrect

Your teacher hands back your paper and you realize you were marked a few points because of a few typos. You had read it the night before turning it in, but did not catch it. What happened?

When looking for the answer to this, I found that one of the main findings of previous literature is that as long as the first and last letter are in the correct place, we can accurately read the sentence or paragraph. For instance:

"Aoccdrnig to a rscheeah at Cmabrigde Uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be at the rghit pclae." 

But, turns out this is not the only important thing to deciphering ambiguous words. When talking about language, we learned how language is a hierarchical structure where phonemes are the smallest unit of words and phrases are the more complex forms of language. This article mainly explains how we actually take into account the sound of the words and the context in which we hear it to make sense of such ambiguous words.Image result for brain autocorrect

This is the first study to specifically look at phoneme ambiguity and how it is influenced by subsequent context to ‘autocorrect’ a word when hearing a sentence. In the first experiment, participants were given two syllables on a screen (e.g. B or P) and then were played the syllable. After hearing it, they were asked to choose which they heard. There was no limit placed on the response time and participants could move on when they were ready. In the second experiment, new participants listened to a word with the word having the same beginning syllable as those in the first experiment. After, they were asked to read the word on a screen and say if it was match or mismatch.

What was different between these two experiments was that the first played the syllable followed by a silence while the second played the syllable followed by the rest of the word. What the researchers were getting at was to see if subsequent context influenced how we interpreted a word as ambiguous or not and how fast it took. Both experiments all had participants wear a magnetoencephalography to record the participant’s auditory cortex.

After looking at the MEG’s of participants, researchers found the following results when it came to sensitivity towards phonological ambiguity in the brain, specifically, the primary auditory cortex:






Figure B shows the location that appears to sensitive to ambiguity in Experiment 1 where just the syllable was given followed by silence. In comparison, Figure D shows the location that appears to be sensitive to ambiguity in Experiment 2 where the syllable is followed by the rest of the word. The results found that phonological ambiguity occurs at an early point (around20 ms after onset). This suggests that there is more sensitivity to ambiguity when given just the syllable (Figure B) versus the word (Figure D) since the word gives more of a context to decipher the ambiguous word versus giving just the syllable and nothing else.

The results further showed that response times were significantly slower for more ambiguous items. The researchers note how the auditory system actively maintains the acoustic signal in the auditory cortex while trying to figure out what the ambiguous word is. For example, if a participant hears the word ‘parakeet’, is given the word ‘barakeet’ and has to determine if it is a match mismatch, it may take a little longer to decide because we have to go back and re-evaluate what we heard before. Then, once we re-evaluate it and realize it is a mismatch do we press the ‘mismatch’ button.

Another explanation researchers talk about as to why there is a slower response time is that as we hear each subsequent sound, we re-evalate the previous phonemes to update interpretations as necessary.

Once part of the method I found interesting was that they allowed participants to take as long as they needed to properly give a response. It might be interesting to see how many mistakes would be made in a time-constrained situation given the finding of the study.

So how does this all relate to autocorrect and the brain? Going back to the example of typos in the paper you turned in for class. When reading the paper, you probably did not read it out loud and because of that, you did not catch the typos. By reading the paper out loud, the sound that reaches the auditory cortex are held while the brain interprets the combination of phonemes to figure out the correct word. This would most likely make it easier for you to catch any mistakes or typos.

1 thought on “Phonological Ambiguity and Autocorrect

  1. kourt21

    When I’m writing information for a class, I noticed that my inner voice goes a lot faster than my hands can type. A lot of the time when I re-read over something I have written or typed, I realized that I skipped over the “the’s” or the articles and connecting words. Not because I didn’t think of them but my hands were trying to keep up with my inner thoughts on what to write. I have seen many study tips online that recommend that students should copy/paste their typed assignments into google translate and have it read it out loud to them. For them to hear the flow of the words and to see if anything is missing.


Comments are closed.