Sound-alikes vs. near sound-alikes
Posted: January 1st, 2017, 1:34 pm
I am beginning to make some good progress into my excursion into orthography, and here is some of my current thinking:
1. The main lexical entry will represent the most common probable spelling based on some metric. Alternative forms of the word are then linked to the lexical entry. This is already done, but I will have to change some of the choices made for main lexical entries after I apply the metric.
2. Alternative spellings of a word are then linked to the correct form of the lexical entry. I have already done this with the help of a program that finds homophones by breaking the words down into phonemes and then applying a set of rules.
3. Any words that don’t follow the rules in category #2 are then considered to be misspellings and are also linked to the correct form of the lexical entry and classified by the type of error i.e. dittography, haplography, etc. (thanks to Stephen Hughes for the idea of classifying them). This is also almost finished.
4. And then there are other types of errors where it is unknown what word is trying to be represented by various stray letters.
My current concern lies in category 2, as it is currently a big category and there is quite a continuum between what may be considered an alternative spelling versus what may be considered a misspelling. So I am trying to see if I could break down step 2 into two categories: 2a) sound-alikes and 2b) near sound-alikes.
Rules that perhaps I would use as sound-alikes for category 2a are:
e=ai, i=ei, o=w, u=oi,
g=c, k=c, d=q, d=t, q=t, z=s, m=n, p=f,
gx=x, mp=p, my=y, nb=mb, ng=gg, nk=gk, nl=ll, nm=mm, nx=gx, np=mp, nc=gc, nf=mf, ny=my, ns=s, nz=z
a=aa, b=bb, d=dd, g=gg, q=qq, k=kk, l=ll, m=mm, n=nn, p=pp, r=rr, s=ss, t=tt,
And then any leftover rules would make up near sound-alikes in category 2b, but I am not very sure about a lot of them:
a=e, ai=ei, a=au, a=o, a=w, u=ou, oi=ou, h=e, h=i, h=ei, i=e, u=i, e=ee, g=k,
(My apologies for not using unicode here, I just cut and pasted them out of my program.) Certainly, these substitutions are being made by various scribes, but I guess my problem is I don’t know how much something sounded like something else. Being originally taught Erasmian pronunciation is not helping at all, as it is hard to see how some of those sounds could be confused for others. Why would “α” ever sound like “ω”, or “α” like “ε”, or “ε” like “ι”? Did “γ” not make a distinct enough sound from “κ”? Buth said that “η” was not pronounced consistently so perhaps it belongs in category 2b. I am also not too sure about some of the softenings of consonants made in category 2a. Is some of this driven by dialect? How would you apply these rules? Or would you even make a distinction between 2a and 2b? I am looking for any input you can give me here.
1. The main lexical entry will represent the most common probable spelling based on some metric. Alternative forms of the word are then linked to the lexical entry. This is already done, but I will have to change some of the choices made for main lexical entries after I apply the metric.
2. Alternative spellings of a word are then linked to the correct form of the lexical entry. I have already done this with the help of a program that finds homophones by breaking the words down into phonemes and then applying a set of rules.
3. Any words that don’t follow the rules in category #2 are then considered to be misspellings and are also linked to the correct form of the lexical entry and classified by the type of error i.e. dittography, haplography, etc. (thanks to Stephen Hughes for the idea of classifying them). This is also almost finished.
4. And then there are other types of errors where it is unknown what word is trying to be represented by various stray letters.
My current concern lies in category 2, as it is currently a big category and there is quite a continuum between what may be considered an alternative spelling versus what may be considered a misspelling. So I am trying to see if I could break down step 2 into two categories: 2a) sound-alikes and 2b) near sound-alikes.
Rules that perhaps I would use as sound-alikes for category 2a are:
e=ai, i=ei, o=w, u=oi,
g=c, k=c, d=q, d=t, q=t, z=s, m=n, p=f,
gx=x, mp=p, my=y, nb=mb, ng=gg, nk=gk, nl=ll, nm=mm, nx=gx, np=mp, nc=gc, nf=mf, ny=my, ns=s, nz=z
a=aa, b=bb, d=dd, g=gg, q=qq, k=kk, l=ll, m=mm, n=nn, p=pp, r=rr, s=ss, t=tt,
And then any leftover rules would make up near sound-alikes in category 2b, but I am not very sure about a lot of them:
a=e, ai=ei, a=au, a=o, a=w, u=ou, oi=ou, h=e, h=i, h=ei, i=e, u=i, e=ee, g=k,
(My apologies for not using unicode here, I just cut and pasted them out of my program.) Certainly, these substitutions are being made by various scribes, but I guess my problem is I don’t know how much something sounded like something else. Being originally taught Erasmian pronunciation is not helping at all, as it is hard to see how some of those sounds could be confused for others. Why would “α” ever sound like “ω”, or “α” like “ε”, or “ε” like “ι”? Did “γ” not make a distinct enough sound from “κ”? Buth said that “η” was not pronounced consistently so perhaps it belongs in category 2b. I am also not too sure about some of the softenings of consonants made in category 2a. Is some of this driven by dialect? How would you apply these rules? Or would you even make a distinction between 2a and 2b? I am looking for any input you can give me here.