Week 8 Update

This week, Dr. Ingram had me do a reliability check on all my data so far. This is basically fixing all the errors in my spreadsheets, which takes a really long time and is pretty tedious. He pointed out some mistakes that I was making consistently while typing up the word shapes that were causing my PMLU calculations to be kind of off. Just to remind you guys, I’m using the PMLU values as a measure of word complexity. For a refresher on what exactly the PMLU values are and we figure out the shape of the word, check out my week five update (I think? I’d check but I don’t have access to the Internet right now. I’ll fix it later. If you guys are reading this, it means I didn’t proofread this post. Whoops.) Besides the random human errors that happened while I was typing things in, there were three main types of mistakes that I was making in my word shape transcriptions. First, I was considering ‘er’ as VC (vowel followed by consonant) instead of V (just vowel) like I was supposed to. This seems kind of confusing at first, but if you think about it, when you say a word like ‘teacher’, the ‘r’ at the end changes how you say the ‘e’ that precedes it, but it doesn’t really stand alone as its own unit of sound. Second, I was counting each complex vowel, or diphthong, as multiple vowels when I was supposed to count them as only one.  For example, the vowel unit in the word ‘boy’ is a dipthong, consisting of two sounds, the ‘o’ as in bowl and the ‘ee’ as in sheep. Even though there are two sounds, the shape of the word ‘boy’ should still be written as CV (one consonant and one vowel). Third, I was counting the consonant sounds ‘ch’ as in cheese and ‘j’ as in judge as two consonants instead of one. I thought that because the IPA transcription of these sounds had two symbols, I should consider them as two consonants each, but this is actually not true.

To be honest, none of these mistakes are very likely to affect the total average PMLUs of any categories. Since I made the errors consistently, they re probably equally likely to be in either the produced or comprehended category. If not, that's actually an important part of my results, because that means certain aspects of words are prevalent in a certain category.

Next week I’m supposed to do the same thing for my semantic categories, which is going to be pretty interesting I guess.