Hi,
I am currently doing a linguistic research and using TAALES (version 2.2) as one of tools for analyzing my text data set. According to the result of my models, two TAALES features, OG_N and OG_N_H could be significant attributes for the label.
However, while I tried to explain why these features contribute to the models, their definitions really make me confused. As the information you provided in the index description, they mean 'phonographic neighbors', one included homophone and another one excluded. Moreover, 'phonographic neighbors' means two words differ in one letter and one phoneme, which have different pronunciation, and homophone means two words have the same pronunciation. Then my question is, for 'phonographic neighbors (homophone included)', how can two words exist with the same pronunciation (homophone) but requiring them differ in one phoneme (phonographic)? I think these two concepts are in complete conflict.....
In addition, while I was constructing feature set using TAALES, some of features were generated exactly the same values, but they have completely different definitions. For example, as the figure shown below,
OG_N_H (Phonographic Neighbors (homophones excluded))
and Freq_N_OGH (Phonographic Neighborhood Frequency Logarithm (homophones excluded))
. May I know what might be causing this?

Looking forward to your reply! : )