Talk:Indo-European sound laws

I wonder if there isn't some better way of formatting this. For me, the vowel chart extends so far to the right I have to scroll the page horizontally to see it all. --Angr 15:29, 24 January 2006 (UTC)

Hittite oi?
Hittite doesn't use the /oi/, in fact, it doesn't use /o/ vowel at all.

- 82.139.47.117 21:03, 20 May 2006 (UTC)

Tocharian fixes
Tocharian only uses ś not š. I also added the retroflex s as an outcome for PIE *s- Imperial78

Merging cells
Does anyone else think that it would be more informativ if adjacent cells with identical information were merged? Everything with [b] for *b, say. A bit of column juggling could be combined with that, altho that might then risk listing unrelated developments (eg Armenian and Germanic [p] for the previous?) under a single entry, which would be counterproductiv. --Tropylium 13:16, 1 May 2007 (UTC)

*n > OCS n
The fact PIE *n has only developed into Old Church Slavonic n seems strange for me. Compare German kinder-like thing and Russian чадо /čado/ < /*kęd-o/ < /*kind-/ (an archaic word for a kid). So, after vowels and not before vowels it has disappeared causing a nasalisation of a previous vowel. Andrew Trevor (talk) 13:59, 28 January 2008 (UTC)
 * Yeah, PIE *n only remains n before vowels in Slavic; otherwise it disappears with nasalization of the preceding vowel. I think PIE syllabic *n also became a nasalized vowel. —Angr If you've written a quality article... 17:28, 28 January 2008 (UTC)

Suggest a non-tabular format for this article
Grover cleveland (talk) 04:15, 15 September 2008 (UTC)
 * The tables are way too large in all directions for my browser
 * The column headings very soon go off the screen so that I can't see what languages I'm looking at. For example, it took me a while to figure out what happens to PIE "m" in Latin.  This negates any advantage of the table format.
 * This would also allow prose discussion where necessary, instead of hiving off all-too-brief mentions to the footnotes
 * We could add examples.
 * It would also allow for sources, something that seems to be missing but is required by WP:V

I think there should be both; a simple table, then a list of notes. --Nathan M. Swan (talk) 18:38, 14 December 2010 (UTC)

*w > Greek h
I can't think of any examples of Indo-European w- changing to Greek h-. Is there a source for this claim that gives examples? It is true that sw- changes to h- (*swādus → hēdys "sweet") but that is development of s, not w. Perhaps the claim came from the Proto-Greek language article, which I corrected afterwards. — Eru·tuon 16:50, 20 January 2010 (UTC)


 * Isn't ἕσπερος related to Latin vesper? --ἀνυπόδητος (talk) 17:38, 20 January 2010 (UTC)
 * Yes, and Andrew Sihler in A New Comparative Grammar of Greek and Latin provides some other examples: ἵστωρ 'wise man' < *wid-tor-, ἕννυμι 'I clothe' < *wes-, Ἑστία = Vesta, ἑκών 'willing' < *weḱ-, ἕδνα 'wedding gifts' < *wed(h)-. No one seems to know when *w- became h- in Greek and when it didn't. The fact that *wid-tor- shows w > h while all the other forms of *weid- don't shows that it isn't simply a matter of *Hw- becoming h- while plain w- became zero (which is the usual explanation for why *y- sometimes became h- and sometimes became z-, namely that *Hy- became h- and plain *y- became z-). +Angr 18:15, 20 January 2010 (UTC)


 * Aha, thanks. Those are all examples I've encountered, but somehow didn't recall. I'll re-add this to the Proto-Greek page too when I have time. — Eru·tuon 14:08, 21 January 2010 (UTC)

Acute for palatals?
I changed palatals from circumflexed k and g to acute-accented k and g to follow the other articles (and in order to have a pre-combined character), but perhaps the circumflex is intentional, so that there's no confusion between acute and glottal-theory apostrophe in fonts where those are similar. If so, feel free to revert my edit. — Eru·tuon 19:15, 22 January 2010 (UTC)

Voiceless aspirates in Sanskrit
Hi! After reading this article, I was wondering where the voiceless aspirated plosives of Sanskrit came from. They are not listed as the outcome of any PIE phoneme in this table, nor in the one at Proto-Indo-Iranian language. However, at the talk page of the latter, User:Grover Cleveland says they are hypothesised to come from a sequence plosive + laryngeal, without citing any source. Do you know anything more about this, or do you perhaps have sources for this, because I feel it is worth mentioning in the article? JaneStillman (talk) 16:28, 26 September 2010 (UTC)
 * Yes; at least some voiceless aspirates of Sanskrit come from voiceless stops followed by a laryngeal, e.g. tiṣṭhati < ; pr̥thu < . Other voiceless aspirates seem to come from s mobile followed by a voiced aspirated stop, e.g. skhalati < *. This is discussed in R. S. P. Beekes, Comparative Indo-European Linguistics, p. 132 & 144. —Angr (talk) 20:39, 26 September 2010 (UTC)
 * This kind of a thing should probably be detailed in the Vedic Sanskrit article, which is however currently rather on the slim side viz. changes from PII. (Unless we want a separate Proto-Indo-Aryan article, a la Proto-Greek?) -- Trɔpʏliʊm • blah 14:27, 29 September 2010 (UTC)

Draft
See this unfinished draft. I don't have any information, I just copied from the table. People who can help? --Nathan M. Swan (talk) 22:43, 14 December 2010 (UTC)


 * Hmmm… grouping might be a good idea. Perhaps a basic division into stops (perhaps secondarily by phonation × POA and only describing individual consonant changes when they deviate from the line), *s, nasals, liquids, semivowels, basic vowels, diphthongs, syllabic nasals, syllabic liquids. Just brainstorming though. -- Trɔpʏliʊm • blah 03:08, 21 March 2013 (UTC)

*p → kʷ before kʷ in following syllable for Latin
In table this change is not showed. Examples:
 * PIE *penkʷe 'five' → Latin quinque;
 * PIE *perkʷu- 'oak' → Latin quercus;
 * PIE *pekʷ- 'cook' → Latin coquere;91.76.135.242 (talk) 17:20, 20 March 2013 (UTC)
 * There is a high number of conditional changes like this that are not shown. Attempting to cover the complete descent of all Indo-European languages with a single table-based approach like this would almost certainly end up too complex and footnote-ridden to be useful. -- Trɔpʏliʊm • blah 21:38, 20 March 2013 (UTC)

Guidelines for soundlaw inclusion
Expanding on the previous IP query, it is a good question though on where to draw the line in including stuff here rather than in the language-specific articles?

My opinion is that the main table should show
 * 1) the main, unconditional reflexes of PIE segments in its descendants
 * 2) any phoneme splits that create new segments common to a main IE subgroup, and can be described in terms of a PIE environment. E.g. palatalization of velars in II, Albanian, Tocharian etc.
 * 3) conditional mergers when they applied to "instable" structures that were widely across IE changed (possibly in differing directions); e.g. the treatment of as *t+t, final *m, perhaps *t+k, *ḱ+w.

*p → *kʷ meanwhile is an Italo-Celtic isogloss and would not qualify. I also think the *s + plosive series, which is widely reflected without any specific changes, and the Greek palatalization should go out.

I also support adding text sections as suggested by NMS. PIE → "immediate" daughter (Proto-Italic, Proto-Indo-Iranian, Proto-Balto-Slavic, etc.) changes could be added here more liberally. I'd still shy away from attempting to cover later developments though (specific to e.g. PC → Welsh or PBS → Lithuanian), short of noting some general changes useful for illuminating things with early-attested examples. E.g. Attic Greek *ā → ē could fall under this.

A section on early PIE → late PIE could also be useful. Other than that I do not support by-language grouped sections; these belong in the individual articles such as Proto-Greek, Proto-Germanic. -- Trɔpʏliʊm • blah 22:15, 20 March 2013 (UTC)

V and Z
There are 2 sounds that English words can start with, V and Z, that do not appear in the English column of the consonant table. Where would they go?? Georgia guy (talk) 22:32, 25 June 2013 (UTC)
 * V appears in the *p and *bh rows. Voiced S can go back to *s, but I can't think of any native English word where this sound is spelled Z (except maybe "haze/hazy"). Most V's and Z's in English appear in loanwords from Franch/Latin. --ἀνυπόδητος (talk) 09:25, 26 June 2013 (UTC)
 * "crazy, maze, hazel" are other (tentative) examples. Maybe we should add Z to the *s row? --ἀνυπόδητος (talk) 09:28, 26 June 2013 (UTC)
 * But those are within a word. At the beginning of a word, *p is F in English, and *bh is B. Georgia guy (talk) 14:05, 26 June 2013 (UTC)
 * Oh. I missed that you wrote "...that English words can start with". I can't think of any native English words starting with V or Z. Have a look at etymonline:V and etymonline:Z and see whether you can find any :-)
 * --ἀνυπόδητος (talk) 15:17, 26 June 2013 (UTC)
 * I know of vat and vixen, acquired into standard English from an SW English dialect featuring the general initial voicing of *f *θ *s that is also found in Dutch and some varieties of German. (More widespread in English is voicing in the, then etc. which is due to prosodical reasons.)
 * That said, this kind of thing happens quite widely (Armenian is a particularly severe case, I believe) and the topic cannot be treated in detail in this article — you're better off asking questions like this under the individual "History of $LANGUAGE" articles. -- Trɔpʏliʊm • blah 20:56, 26 June 2013 (UTC)

Another comment still: All word in Latin that have z in them are loanwords from Greek. This means that any loanword from Latin with Z is in fact a loanword from Greek. (This is a response to Anypodetos's statement above that says "Most V's and Z's in English appear in loanwords from French/Latin.") Georgia guy (talk) 19:18, 26 June 2013 (UTC)
 * The only native words starting with V that I know of are vat, vane, and vixen. In all three cases, the v comes from Old English f by a regular sound rule that applied only to some dialects of southwestern England. For whatever reason, those three words entered the standard language from a dialect that had that voicing rule, while most other words that started with f in OE entered the standard language from other dialects without the voicing rule, so we say four and father and fire rather than *vour and *vater and *vire. Those same dialects also voiced s to /z/ word-initially, but apparently no words with initial /z/ entered the standard language from those dialects. Angr (talk) 20:58, 26 June 2013 (UTC)

Greek plus j
What does Greek plus j mean?? Georgia guy (talk) 23:53, 28 June 2013 (UTC)
 * It means that column shows the Greek reflex of the respective row's consonant when followed by j (i.e. PIE *y/i̯). For example, *pj becomes pt in Greek. Angr (talk) 09:47, 29 June 2013 (UTC)

C and K vs. Ch
The table says that PIE initial g yields either c or k in English; c in general but k before e or i. But then it also says ch, with a 10 superscript that must mean before front vowels, which I believe means e or i. Any corrections?? Georgia guy (talk) 14:46, 26 July 2013 (UTC)
 * It became ch before Proto-Germanic front vowels, as in child and chin; it became k before e or i that came from Proto-Germanic back vowels, as in king and keen. Aɴɢʀ (talk) 06:57, 27 July 2013 (UTC)
 * Can you explain the difference between the words chin and king when it comes to vowel etymology?? They have the same vowel. Georgia guy (talk) 18:10, 29 July 2013 (UTC)
 * Chin comes from Proto-Germanic *kinnuz; the /k/ became /tʃ/ before /i/ in the history of English (but not in German, which has Kinn). King comes from Proto-Germanic *kuningaz where the vowel after the /k/ was back so it didn't become /tʃ/. Later on, the /u/ became /y/ by umlaut, and later still that /y/ became /i/, but by that time the rule changing /k/ to /tʃ/ before /i/ had stopped applying. Aɴɢʀ (talk) 19:06, 29 July 2013 (UTC)

*t+t [tst]
I'm working on a personal project involving PIE and I'm using the table of this article as reference. I've noticed lots of irregularities between some information in the table, but the one that concerns me most is the "*t+t" row, I haven't seen a PIE word that has either *-tt- nor *-tst-, why is this? Another of my worries are the germinated vocalic consonants. — Preceding unsigned comment added by 201.172.36.167 (talk) 04:14, 28 July 2013 (UTC)
 * t+t occurs mostly if not only at morpheme boundaries, such as when a root ending in t is followed by a suffix starting with t. I don't think PIE had any geminate consonants; even t+t at a morpheme boundary resolved into [tst], and s+s at a morpheme boundary was simplified to plain s (e.g. *h1esi 'you (sg.) are' < *h1es- + -si). Aɴɢʀ (talk) 08:44, 28 July 2013 (UTC)

Gwh in English
Why did word-initial gwh in PIE yield b, as opposed to gu pronounced gw?? Georgia guy (talk) 15:29, 28 August 2013 (UTC)
 * Well, because word-initial PIE *gʷʰ became *b in Proto-Germanic, which stayed unchanged in English, though that answer just shifts the question elsewhere. It's not usually possible to say why a sound change happened at all (why did Grimm's Law happen? why did Verner's Law happen?). It's just one of those things. For whatever reason, *kʷ and *gʷ remained labiovelars in Germanic, while *gʷʰ lost its velar element and became a labial. For what it's worth, the exact same thing happened in Celtic: *kʷ and *gʷ remained labiovelar in Celtic, while *gʷʰ lost its velar element and became *b. Aɴɢʀ (talk) 15:44, 28 August 2013 (UTC)

Map
The isoglosses map is ugly, confusing and not as informative as it could be.

What would perhaps be better would be to have one outline for the total extent of the IE family, or perhaps a shading for those areas outside it, and within the IE area use different, clearer colours with subtle shading / hatching depending on what works, and denote the isoglosses somehow in colour in the map. — Preceding unsigned comment added by 82.139.82.82 (talk) 00:56, 9 July 2015 (UTC)

Persian?
I know that there's an Avestan column, but would anyone object to a column for modern Persian? — Preceding unsigned comment added by Fearfuljesuit (talk • contribs) 12:54, 11 November 2015 (UTC)
 * The conventions here so far would suggest Old Persian being a better idea. Trying to cram in all hundreds of modern IE languages is obviously unfeasible, and trying to pick "significant enough" representatives seems unlikely to work either. Currently we only have English, what with this being the English Wikipedia, and Lithuanian & Albanian, for which ancient representatives are not available.
 * (I also suspect that readability would be improved by switching to listing reflexes instead in e.g. Proto-Iranian or Proto-Germanic rather than in attested languages, but my impression is that the intermediate IE protolangs are mostly not in a good enough shape yet for this to be really doable.) -- Trɔpʏliʊm • blah 13:23, 13 November 2015 (UTC)

Cluster table
Possibly giving the usual reflexes in grayed-out would be productive. Any opinions? -- Trɔpʏliʊm • blah 14:10, 27 August 2016 (UTC)

Can someone add in the kʷ subsection
The source for all other example words just so happens to not have the pages containing *kʷ to be shown in a preview, can someone with access to the book add the subsection in? Henry Wong ts (talk) 14:35, 11 October 2020 (UTC)

Inclusion of Brittonic?
The table of changes should include Welsh as well as Irish (if not instead of) for the Celtic branch. Welsh is the healthiest of the Celtic languages and isn't even represented - it should sit with Irish to give clear comparisons between the Brittonic and Goidelic branches. - Dyolf87 (talk) 10:23, 13 December 2022 (UTC)

If sb. could please add an annotation that C + /j/ means consonant + j. Thanks!
If sb. could please add an annotation that C + /j/ means consonant + j. Thanks! The table should be comprehensible for readers without such prior knowledge. 92.184.124.136 (talk) 21:43, 3 April 2023 (UTC)


 * I added a note to both of the contexts where C shows up on this table. Hope that helps. Cheers! JungleEntity (talk) 01:13, 4 April 2023 (UTC)

*Proto-Germanic instead of English/Gothic
I can see that the idea of the table is to list only languages that are not reconstructed, as "target" languages.

However I believe it would be more helpful to remove "English / Gothic" and replace it with just "Proto-Germanic" (adding a * in front of consonants of course). Subsequently, a separate table for Proto-Germanic could list reasonably-chosen descendants of Proto-Germanic, like this: PIE ProtoGermanic Old English Old High German Old Saxon

and then a table for German: PIE ProtoGermanic Old High German, Middle High German, Modern Standard German

Likewise, for Latin:

PIE Latin Old French Old Spanish Old ... etc.

and for French: PIE Latin Old French Middle French Modern French.

A special table for Anglo-Norman sound shifts would be helpful, too, even though these is of course adaptations from a Romance language into Germanic.

PIE Latin Anglo-Norman Middle English Modern English

In short, I find the current table difficult to use because it lacks important junction points from which important languages developed (Proto-Germanic -> German, Englisch, Norwegian, ...), and because there are no tables that make use of this great overview by expanding it further for languages that underwent considerable change, such as German and French, whilst retaining the important reference to the original PIE, i.e.: PIE Latin Old French Middle MIddle French Modern French

The table form (as well as retaining PIE, as well as Proto-Germanic (resp. Latin) as anchor points) is very helpful, unlike endless lists of bullet points in some articles covering "Middle French" etc. 80.187.115.179 (talk) 04:02, 6 April 2023 (UTC)

PIE ḱ in albanian
Could someone please explain how and why it evolved into a th sound in albanian? What went differently for the sound to not evolve into a s or š sound like other satem languages? An answer would be highly appreciated. Rigers15 (talk) 22:31, 17 April 2023 (UTC)

Color-coding the tables
I have undertaken some effort color-coding the first table; in order to do this for the WP article, it would be necessary to place every exception (that is denoted with a C ## reference) into its own cell within the table. (Note: Since non-referenced consontants has been chosen to be non non-referenced for a reason ("defaults"), it would make sense to make sure their colored area is larger, and possibly (at least partially) forms a "frame" around the non-default consonants.)

As colors, I have chosen:

p -> blue b -> light blue v -> blue-violet f -> violet

t -> red th -> medium red d -> light red

k -> green g -> light green x -> intense green h -> yellow

s -> orange r -> brown

It's very interesting to see patterns emerge thanks to the coloring. I just wanted to share this in case sb. has been thinking of doing the same, and in case sb. wants to improve the table by using colors. 77.181.88.229 (talk) 14:04, 19 April 2023 (UTC)