Talk:Proto-Germanic/Archive 1

From Wikipedia, the free encyclopedia

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Contents 1 proto-germanic people? 2 edits 3 shouldn't? 4 Value of /d/ 5 Hard to follow text concerning the origins of Proto-Germanic 6 Tenses 7 Cases 8 Nominal Declension 9 The opening phrase ... 10 phonology/notation 11 An apparently mojibake-ified symbol 12 Meaning of words 13 Grimm's Law Question 14 Kündü Ghur proto germaic? 15 The Voiced Velar Nasal 16 Intervocalic hardening 17 Centum - Satem connection 18 Recent change about origin 19 St <> F 20 Protolanguages are hypothetical and reconstructed 21 Timing of the First Sound Shift 22 Personal pronouns 23 The Map 24 Requested move 25 Notes and references vs. References 26 Creole Shmeole 26.1 Hybridization theory 27 Some excisions from the Wikipedia Proto-Germanic article

proto-germanic people?

Since there was a proto-germanic language, then logically there would have also been a proto-germanic people. Gringo300 21:27, 17 July 2005 (UTC)

Probably.--Wiglaf 21:30, 17 July 2005 (UTC)

and maybe a proto-germanic CULTURE. Gringo300 22:14, 3 August 2006 (UTC)

teoreticaly may they have a big country ? :) --"the eldest man under sun" 16:13, 8 August 2006 (UTC)

I don't think it could be claimed they were unified enough to possess a "common country", a common culture seems more likely, based on their religion and such traits... 惑乱 分からん 17:06, 4 November 2006 (UTC)

There would have been a Proto-Germanic speech community but not a Proto-Germanic people. It is highly unlikely any group in antiquity ever identified themselves as such and instead were more likely to identify themselves by their individual tribal/socio-political affiliations. Trollaxor 12:48, 20 September 2007 (UTC)

edits

Ken Alan's edits raise a general point worth discussing: there are fields where a lot of scientific progress has been made, where knowledge and training are needed to participate competently, and where there exists a periphery of amateur participants who hold very different views. Historical linguistics is clearly such a field; evolutionary biology is arguably another.

The question is how Wikipedia editors, who may have some training in a field but are usually not experts in it, should deal with such situations. How can mainstream editors ensure that the Wikipedia is a professional and competent product, without seeming to violate the principle of NPOV?

I think the answer lies in a point made already in Wikipedia:No original research, namely that we should not be putting our own thoughts on the Wikipedia at all, but rather doing our best to report what credentialed, knowledgeable scholars have found and published in books and articles. It's true that this is not always easy to do in practice (and I have occasionally violated the principle myself, mea culpa), but it seems the safest way to make sure that our product is of high quality.

Opus33 01:21, 12 Apr 2004 (UTC)

It may be one of the few cases of ethnic origins were there is a relative consensus. Even, though, I am not a specialist in the field, I have taken the liberty of defining the original homeland of the Germanic tribes as corresponding to the Nordic Bronze Age, which also fits well with linguistic evidence. I feel the that "the plains of Denmark, southern Sweden and northwestern Germany is a much clearer definition than "west Baltic" which may cause a confusion between the Germanic tribes and the Baltic tribes. Wiglaf

As an archaeologist, I feel extremely uncomfortable with the equation archaeological culture/language (or even more 'proto-language'). As far as I know, there is no evidence of any immigration at the beginning of the Bronze Age (Unetice?) in Northern Germany. Gimbutas has claimed evidence for an invasion of the bearers of the Corded ware culture (late Neolithic), but this is extremely shaky as well.

--Yak 10:14, May 10, 2004 (UTC)

Yes, I know that most archaeologists are uncomfortable with that equation! When we lack a written language it is impossible to say what language a particular group of people spoke. However, the Nordic Bronze Age and its evolution into Germanic Iron Age, fits the image of a continuous Germanic culture AND language.

The problem with most archaeologist is that they usually know very little about linguistics, and language change, and language acquisition and so avoid the subject. I can assure you that the cases of peaceful language displacement or language change without migration, which some archaeologist hypothesize about, is extremely rare. The only cases I know of are a few pidgin language that have evolved into geographically limited Creole languages. We cannot exclude that the Germanic languages arose as such a language during the bronze age by trading with an Indo-European trade empire. However, the structure of the Germanic languages have very little in common with the structure of creole languages. Moreover, for such a language to be as homogeneous as the Germanic languages were would be a spectacular feat indeed. What is relevant in this article, however, is whether the proto-Germanic language is connected with the Nordic Bronze Age. I, personally, bet it is, and I have not met any hypotheses that contradict it. Wiglaf

Well, even a PhD in linguistics has to be able to present proof that whatever he writes in a wikipedia-article is widely accepted in the relevant fields. The rule on original research is valid here too, as long as any personal theories have not been published in a peer-reviewed scientific journal. But I assume that it will be no problem at all for you to assemble a list of references validating your views, or is it? By the way, just out of sheer curiosity, what is the subject of your PhD that you are working on and in what research group at the University of Uppsala? Fedor 20:04, 10 Jun 2004 (UTC)

I'd rather keep personal information to myself, thank you very much. Here you have a reasonable site with sensible information, although speculative concerning the Fenno-Ugric connection. It has a bibliography to ease you curiosity: http://www2.4dcomm.com/millenia/hunters.html regards.--Wiglaf 12:35, 28 Jul 2004 (UTC)

Battle axe: late Neolithic. This seems to be the aged Gimbutas-Stuff. Why don't you give the source? --Yak 19:26, 9 Aug 2004 (UTC)

I just gave you one. Moreover, Gimbutas' theory may be old, but I am still waiting for theory that explains how languages are spread without migration. I know a few archaeologists in person who believe that I-E languages could have spread by "cultural diffusion". However, they don't know squat about languages. People don't change languages like they change tools. A new language is hard to learn for most people. Morevover it is inherent to a person's social identity and they only change if it is to their social advantage. There are many historical examples of language change. Not a single one has happened without the pysical establishment of a power structure that favoured the new language. If you want to discover the naivité in believing that language change can happen otherwise you can have a look at this relatively new dissertation [1].--Wiglaf 20:14, 20 Aug 2004 (UTC)

Cultural diffusion can include exchanging women as wives over tribal boundaries. The new wives can then raise their children speaking (at least partially) the woman's native language. Over an extended period of economic stability, this can lead to large scale linguistic change. We are talking about hundreds, even thousands of years here. So, I thought to put at least a mention of it in the article. Fire Star 01:30, 24 Aug 2004 (UTC)

Yes, but women use the langage of the men and of the people, where they settle. For information, check the dissertation I mentioned [2].--Wiglaf 05:08, 24 Aug 2004 (UTC)

Greetings Wiglaf. Unfortunately, my browser won't open the links you provided. I'll keep trying. (Hah! It opened, third time's the charm! I look forward to reading it.Fire Star) I agree that typically a woman married (or traded) to or captured by another tribe speaking another dialect would probably want to try to use the language of her new home as well as she could, but as you say, a new language is difficult to learn, so there will be artefact which would accumulate over time. As well, people use different speech patterns for different levels of formality. In the home or around other women from her (or another) original tribe, she may have been less guarded. I do myself believe that there was some population movement brought about by natural pressures (climate or agricultural degradation or epidemics, esp. notable in the 13th-11th centuries BC) but not as much as some (Gimbutas) would have us believe. There are also going to be terms associated with new food sources or farming methods and "prestige" technologies (seafaring, bronze-working, "beaker" utensils, etc.) which will have crossed tribal boundaries even in the absence of large scale migration. What I feel the present article lacks, though it presents several alternate theories as it does, is the reporting of the sense that we don't really know all that well what was going on back then, especially with an intangible such as language. The best we can come up with are more or less educated guesses based on modern distributions of evidence; archaeological, genetic and linguistic. Many of the guesses are ingenious and compelling and it is a fascinating issue, however. I can only keep an open mind. Fire Star 06:38, 24 Aug 2004 (UTC)

Greetings Fire Star. If I seem to disregard the diffusion theory, it is because, as a language teacher and a linguist, it goes against what I know about languages.

There's an even better book about language and ethnicity. It is written by Robin Dunbar and is called Grooming, Gossip and the Evolution of Language. It is a very interesting book that studies the role of language as a social marker and has as a central thesis the role of language for our social life. In this book, he states that what language is used is depedendent on the males of the group. You find the same kind of information in the dissertation I mentioned. Consequently, languages are spread with groups of males, and groups of males usually had to fight to settle in a new territory. There is a reason why the Caucasus Mountains have many languages - they are easy to defend. --Wiglaf 17:20, 24 Aug 2004 (UTC)

Interesting. The groups of males theory seems to be supported by some recently published DNA studies, especially concerning the yDNA and mDNA of the Basques and the Irish of the Gaeltacht. I'll have to do some digging, but I'll try to find the articles for you. Fire Star 17:38, 24 Aug 2004 (UTC)

I'd love to read them. IIRC, the same evidence has been provided by the English-speaking population of England. The yDNA is Germanic, like the language, while the mDNA is Celtic. The same evidence has been provided by genetic studies on the Saamis. The mDNA is autochtonous whereas the yDNA is Fenno-Ugric, like the language. In areas of Latin America, the yDNA is Spanish, like the language, while the mDNA is Amerindian.--Wiglaf 17:46, 24 Aug 2004 (UTC)

But, according to that theory, the English should be speaking French. Burschik 12:24, 25 Aug 2004 (UTC)

Burschik, why? There were only a few thousand Normans who immigrated to England, while there were one or two million Anglo-Saxons and Scandinavians who remained? The Norman language is a classic example of the appearance of a superstratum. However, according to the "diffusion theory" we should all be speaking French after the popularity of French during Louis XIV.--Wiglaf 15:07, 25 Aug 2004 (UTC)

There are many factors to consider. Why did the Anglo-Saxon's Old English completely displace the Brythonic of the Celts, and yet the Normans (while certainly affecting English vocabulary) then failed to displace English with French? Some say that it was because society by that time was more rigidly stratified and the invaders were less likely to wed native women, others (as above) simply because the numbers of the invaders were insufficient. "Diffusion" in mediaeval society is certainly much different than it would have been in the much more extensive period of pre-literate proto-Germanic culture. The time frame is the biggest factor, and one that is hard to get a handle on due to the lack of records. It seems to me that the earlier Bronze and late Neolithic farmers were therefore affected by many different pressures for linguistic change at various times in different places, an "all of the above" situation. The DNA typing for haplogroups and their diffusion and rates of mutation are interesting in this regard. Here is a link to one article I have found so far [3] Fire Star 18:55, 25 Aug 2004 (UTC)

I agree, we're talking about a long period of time and we know nothing of the details. I still think that historically documented language changes are the best analogy.--Wiglaf 19:47, 26 Aug 2004 (UTC)

BTW, thanks for the link. Interesting!--Wiglaf 20:56, 27 Aug 2004 (UTC)

I just noticed this debate. In terms of whether invaders are able to force their language on those invaded, it depends on various things: the relative numbers especially, the policy of the invaders, to some extent the extent of linguistic homogeneity of those conquered and the difference in cultural sophistication between the two. The Anglo-Saxons came in large numbers, and as a people -- to the extent that it is said that their old homelands were largely vacant for a time; and they kicked the Celts off the land they conquered. The Normans came largely as upper-class landowners -- i.e. not in sufficient numbers, and they did not eject the Anglo-Saxons but were content with maintaining the existing system, as they did not work the land themselves.

In general, language only moves either [a] when people move, or [b] by concerted, long-term political domination of an area; often with an explicit attempt to eradicate the indigenous language. (Cf. French in South France and in Brittany.) Benwing 20:17, 2 August 2005 (UTC)

---

I did not call the method highly speculative, but the "knowledge" obtained. Burschik 08:02, 24 Aug 2004 (UTC)

I see, well yes, the results of the comparative method are hypothetical by definition, but I wouldn't call them "highly speculative".--Wiglaf 17:22, 24 Aug 2004 (UTC)

I would consider "highly speculative" and "hypothetical (with no means of verification)" to be more or less synonomous, but I will not insist on wording. Burschik 10:22, 25 Aug 2004 (UTC)

How much do you know about comparative linguistics? Every single known Indo-European language, dead or alive, has constitued a "means of verification". I suggest that you read Cours de linguistique générale by Saussure. He was one of those who developed the method and in the book you find a nice chapter to read. The chapter is directed at those who consider the forms to be pure speculation.--Wiglaf 15:07, 25 Aug 2004 (UTC)

I have an MA in linguistics. I studied comparative historical linguistics for several years. I also studied logic and the philosophy of science. Maybe that is why our opinions on what is highly speculative differ.

But to make my position more clear. I specifically did not call the comparative method highly speculative, but its results in this particular case, namely the reconstruction of Proto-Germanic. First of all, we do not know whether the language ever existed (sure, it is more than likely, be where is the proof?). Secondly, there are no records of the language, nor is it likely that any will ever be discovered. So, reconstructing Proto-Germanic may be fun, but it isn't science. What's that quote again? "No language ever changed as fast as reconstructed Proto-Indo-European." or something like that. Regards, Burschik 06:07, 26 Aug 2004 (UTC)

What is "Wild speculation" to one scholar is a "realistic model" to another. How would you define what we know about Proto-Germanic? As "wild fantasy" or as an "approximation"?--Wiglaf 19:47, 26 Aug 2004 (UTC)

As an approximation. In fact, I quite agree that the model is probably quite accurate, but we have no means of verification (nor falsification). And this is exactly what makes it highly speculative in my opinion. But let's not blow this out of proportion. I added "highly speculative", you removed it, and I am not insisting that you put it back in again. Burschik 05:57, 27 Aug 2004 (UTC)

Thanks! I see what you mean now :)--Wiglaf 14:38, 27 Aug 2004 (UTC)

shouldn't?

Shouldn't this be at Proto-Germanic language instead? -Branddobbe 10:22, Feb 13, 2005 (UTC)

Value of /d/

The article says:

Evidence is conflicting w.r.t. /d/. It is agreed that /d/ is a stop [d] initially, followed by a nasal, and doubled. Gothic and Old Norse suggest that /d/ was [ð] elsewhere, but the other Germanic languages suggest that /d/ was [d] everywhere, with no evidence of it ever having had a fricative value. Note, then, that Gothic and Old Norse show a symmetrical system where /b/, /d/, /g/ are stops when initial, doubled or post-nasal, and fricatives elsewhere. The reconstructed system of the other (West Germanic) dialects, however, is highly asymmetric (/g/ is mostly fricative, /b/ is part stop, part fricative, and /d/ is entirely stop). Analogy works towards symmetry, and hence the reconstructed West Germanic system is likely to be correct and the symmetric systems of Gothic and Old Norse secondary developments. (An additional argument for this is that early borrowings into Gothic corroborate the initial [ɣ] in Pre-Gothic as in West Germanic.)

Could someone please direct me to the evidence (or the lack)? I understood that it was fricativised in Old English, accounting for things like En. father De. Vater? Is that wrong? Why? (I'm an amateur of the sort described in less than glowing terms at the start of this talk page...) Felix the Cassowary 11:58, 1 August 2005 (UTC)

Actually, the evidence from Gothic, though not without its complexities, is pretty clear. Patterns like hlaifs "bread" dative hlaiba, staþs "place" dative stada, witoþ "law" genitive witodis point directly to the inference that the b, d in such words stood for voiced fricatives which, when devoiced in word-final position, became unambiguously fricative. Note, then, that no such thing happens in sequences like -nd-, -rd- and the like: waurd "word", ald "time", huzd "hoard", gards "yard". As mentioned, the facts aren't as simple as they might be: there's no such alternation of g and h: so dags "day", not ˣdahs, as well as other inconsistencies: daug "avail", which fits with dags, but on the other hand there's mag "is able", with preterite mahta and participle mahts. Further there are forms like hlaib "bread" (accusative), but these are usually or always followed by a word beginning with a voiced consonants, in fact usually a vowel or a resonant (nasal, liquid). Nothing says that Gothic is identical to all the other Germanic languages, but the alignments are pretty broad. The stoppification of Proto-Germanic (and Gothic) [ð] in West Germanic is nothing to wonder at; voiced slit spirants are far less stable diachronically than groove spirants, nor is it odd that [ð] and [þ] would part company in West Germanic. Does that help answer your question? Alsihler 21:18, 7 February 2007 (UTC)

father was OE fæder with /d/; the th appeared later and was by analogy with OE brothor. Better examples are OE heavod > head, OE hlæder > ladder (Ger. haubt < OHG houbit, leiter). Benwing 19:39, 2 August 2005 (UTC)

Ah, thank you. That, I suppose, is obvious if you think about it :) Felix the Cassowary 21:54, 2 August 2005 (UTC)

What about the word weather? How would you analyse that with brother?

No, of course not. And neither are father (and mother!) "analogies" with brother or anything else. It's a Middle English sound law relating to the development of -dr-. Lather is in fact attested in the 16th & 17th cents. for now-standard ladder < OE hlǽder. Standard English is a bit of a dialectal ragbag, with forms like ladder, adder, rudder, murder and the like with forms from south of London; the forms with /ð/ are attested, though, for all or nearly all the affected words, with some evidence on top of that of hyper-corrections (ladder for lather (OE léaþor), as in soapsuds). Alsihler 21:26, 5 February 2007 (UTC)

Hard to follow text concerning the origins of Proto-Germanic

Benwing added the following text for which I need some clarification:

Yet other scholars have asserted that the Germanic branch was the last Indo-European branch to break off, suggesting that it actually underwent the least amount of separate development as compared to other branches.

Is this text about Northern Germany as the Urheimat of Proto-Indo-European? If so, please clarify the position.

"last to break off" is a confused wording, suggesting continuing existence of the "original" off which things could break off. from the pov of every group, the others broke off, until only they were left. It's not like they shook hands like Abraham and Lot, people just settled across the steppe over the generations, with remote settlements being lost at the horizon. The concept is really quite similar to the expanding Universe: Everybody was at the 'center', with others 'breaking off'. All this talk of "central" dialects is misleading. The "center" is wherever an innovation crops up, spreading outwards. This may and will happen at completely unrelated points, creating lots of "centers", one for each innovation. There is no "center" of non-innovation, that's just the null case of nothing happening. "Nothing" can happen anywhere, at the "periphery" as well as in the "center". dab (ᛏ) 20:58, 15 August 2005 (UTC)

It is not clear, however, how much of this represents loss of pre-existing categories as opposed to the failure to develop categories (such as the future tense) that were not in the parent language but appeared as innovations in many of the daughter languages.

This piece of text I don't understand at all. I left a note about this, and maybe it is best discussed here.--Wiglaf 21:24, 14 August 2005 (UTC)

this, however, is unproblematic to the point of being superfluous. Nobody claims PIE had the future, so there is no point in emphasizing that Gmc didn't lose it. They did, however, lose lots of things, beginning with the aspect system. dab (ᛏ) 21:00, 15 August 2005 (UTC)

Yes, and the loss of the aspect system is in my mind a MAJOR change. Swedish students notoriously have a very hard time learning what the aspect system is about, whatever the language. I usually explain that it is Germanic that is deviant. This may however be an old substrate thing, but if there was still mutual intelligibility with a neighbouring language, the change should have spread I reckon. Too bad there are no texts from the Nordic Bronze Age or the Urnfield culture.--Wiglaf 21:07, 15 August 2005 (UTC)

Tenses

I think it should be stated in this article that tense is a very convoluted word in modern grammar. In reality, PIE only ever had 2 tenses, past and present. The other tenses in Latin, Greek, and Sanskrit are actually combinations of tense and aspect. It would be best to say that Germanic lost the concept of aspect rather than saying it lost a few tenses. Though I haven't read up on this, I believe this loss of aspect to be due to Germanic's inability to form compound sentenses (i.e. those with more than one clause.) Languages that don't have compound sentences, such as Japanese and Finnish, also lack either the concept of tense (Finnish) or aspect (Japanese.)

Cases

"The instrumental and vocative can be reconstructed only in the singular"

What about the Old English dative plural of se:, þæ:m next to þa:m? I would reconstruct the first as coming from a Germanic instrumental plural þaimi(z) and dative plural þaimuz.

"the instrumental survives only in the West Germanic languages"

Gothic word þe: 'by which' originally an instrumental.

I have a question about the article, unrelated to this person's previous, but it has to do with cases so I thought I would ask it here. In the article it says that Proto-Germanic only has 6 cases (nominative, accusative, genitive, dative, instrumental, vocative), yet in the table of the definite article reconstructions a locative is used. Did someone put locative in there by mistake (in which the L should be a V and the appropriate order changes should be made) or that not a mistake?Deman7001 06:25, 19 May 2007 (UTC)

See http://www.utexas.edu/cola/centers/lrc/books/pgmc03.html. The 6 cases survive for sure in daughter dialects, and beyond that locative and ablative are reconstructed from PIE continuity with only "sparse remnants" in Germanic languages. --Idda 21:22, 19 May 2007 (UTC)

Nominal Declension

-a- stems, masc.

Singular

N. wulfaz V. wulfe A. wulfan G. wulfasa D. wulfai I. wulfō

Plural

N. wulfōz V. = N. A. wulfanz G. wulfōn D. wulfamiz I. = D.

The opening phrase ...

... reads Proto-Germanic, the proto-language believed by scholars to be the common ancestor of the Germanic languages and is unnecessarily confusing. What is the doubt? I am pretty sure that we consider it a fact that there has existed a language from which the germanic languages evolved. The doubt is about the nature and content of this language. What I think we want to say is: The germanic languages evolved from a single ancestral language, Proto-Germanic. We probably cannot know the details of this language, but some of its features can be inferred with reasonable probability from the features of its descendant languages. --Etxrge 07:34, 1 October 2005 (UTC)

Isn't a proto-language not the actual language, but our best guess at it? I was of the impression that the original common language was Common Germanic, which we cannot know (save a time machine), but our best guess at it is Proto-Germanic. —Felix the Cassowary (ɑe hɪː jɐ) 10:24, 1 October 2005 (UTC)

According to the articles about proto-language and proposed language proto-language refers to the actual language. --Etxrge 12:14, 1 October 2005 (UTC)

Proto-language is the language that was actually spoken. If we had no records of Latin, but had to reconstruct it on the basis of the modern Romance languages, we would have something that differed significantly from the reality, especially if we lacked Romanian to alert us to the existence of a case system. But "Proto-Romance" would be Latin as it was actually spoken, not our reconstruction of it. Copey 2 12:54, 31 August 2006 (UTC)

phonology/notation

How shall we spell the PG reflexes of voiced aspirates? The article is inconsistent at present. The easy way would be just b, d, g, but there is đ -- what is the corresponding glyph to be used for g? ȝ (yogh), ɣ, or is there a similar "small letter g horizontal bar"? And what about "small letter b horizontal bar" dab (ᛏ) 10:18, 6 December 2005 (UTC)

An apparently mojibake-ified symbol

the last line of section Proto-Germanic language#Hybridization as conjectured cause says:-

However, the Germanic languages are commonly classified as Centum languages, because of the words *hund, not !sund ("hundred", ~ centum with guttural fricative according to Grimm's law) and *hwis, not !his ("who", ~ Latin quis).

What should the exclamation marks be? (See mojibake.) Anthony Appleyard 07:01, 6 January 2006 (UTC)

I think the exclamation marks are intended to be exclamation marks, but should be asterisks. I think that was written by someone computer-literate who interpreted ! as meaning "not". Of course, the word "not" means "not", and the exclamation mark represents a hypothesised or incorrect form, which is precisely what this is. I'll correct it. —Felix the Cassowary 13:06, 6 January 2006 (UTC)

I corrected it to one asterisk = "hypothesized form", two asterisks = "wrong form", as is usual in linguisitics. Anthony Appleyard 17:16, 6 January 2006 (UTC)

Meaning of words

What do the reconstructed sample words in the tables under "Morphology" actually mean? (I'm talking about *wulfaz, *blindaz, and *gastiz)--ikiroid | (talk) 00:24, 11 March 2006 (UTC)

As you'd probably expect: wolf, blind, and guest. The last one is articles. Compare old english se, seo, and þæt.

Grimm's Law Question

Is it just a coincidence that the Grimm's Law consonant shift of /p/ to /f/ seems to point to an early Germanic form of the number 5 which is essentially identical to the P-Celtic form? This may (or may not?) support the hybridization hypothesis discussed in the article, but there may be other explantions, as well.

Also, it was odd that the hybridization hypothesis mentioned the possibility of influence by a centum, a satem, and a non-Indo-European language, but without noting that the article's maps suggesting possibilities for proto-Germanic-speaking areas do generally show a territory bordering areas where there is every reason to believe there were speakers of a Celtic languages/dialects (centum) to the west, Baltic languages/dialects (satem) to the east and south, and one or more Finno-Ugric languages, which are non-Indo-European, to the east and north. The omission was somewhat striking.

-- Bob

As far as I know, the answer to your first question is yes, it's just a coincidence. PIE "five" has to have been *penk^we, but all three major western European branches assimilated the consonants to each other in one way or another. PGmc *fimf points to *pempe, while both Italic and Celtic point to *k^wenk^we. The Italic and Celtic shift may have been encouraged by the initial k^w- of "four", since adjacent numbers often tend to alliterate with each other. Alliteration with "five" is presumably also the reason why PGmc *fedwar- "four" starts with *f- < *p- rather than *hw- < *k^w-, and alliteration with "ten" is definitely the reason why the Balto-Slavic word for "nine" starts with d- rather than n-. Angr (t • c) 22:20, 16 May 2006 (UTC)

The subtext of your comment appears to be a nonacceptance of what is described in the article as a hybridzation hypothesis for the origen of proto-Germanic. This I cannot take a position on, one way or another. I note, however, that the hypbrization hypothesis may explain "fedwar" and "fimf," as the article on the Gaulish language has "petuarios" as the ordinal form of 4 and "pinpetos" as the ordinal form of 5 in Gaulish. According to the article on Celtic languages, Gaulish is a P-Celtic language, and "[t]he difference between P[-Celtic] and Q-[-Celtic] languages is the treatment of Proto-Celtic *kw, which became *p in the P-Celtic languages...."

--Bob

I had never even heard of the "hybridization hypothesis" until you mentioned, and there are no sources provided, so it may be original research on the part of whoever added it. Considering there are plausible pragmatic and phonological reasons for Germanic and "P-Celtic" (which incidentally I don't believe in either) to have arrived at *pemp- separately, there's no particular reason to believe the Germanic form is due to Celtic influence. It's also extremely likely that Germanic had already gone through Grimm's law, producing *fimf-, long before Gaulish underwent *k^w to *p. In other words, (pre-)Germanic *pempe and (pre-)Gaulish *pempe were never contemporaneous. There was in all likelihood a period of at least several hundred years when Germanic already had *fimf but all forms of Celtic still had *k^wenk^we. Angr (t • c) 20:49, 17 May 2006 (UTC)

Okay, that makes sense. While I did not think that Gaulish was contemporaneous with the Grimm's Law shift, it was never clear to me whether the P- forms in Celtic entirely post-dated the Grimm's Law shift. (You implicitly say that they did.) What is the date of the emergence of these P- forms in those Celtic languages that have them? I had thought it might be early, because I have some recollection that there was also a less-pronounced but arguably similar split in the Italic languages, with one cognate of the Latin "quattuor" in another Italic language being something like "*pattuor," though I do not recall the details. Is there any reason to believe that there was an early tendency to shift between *k^w and *p which could have manifested itself in both Celtic and Italic (though more strongly in the former)? Finally, if the hybridization hypothesis is entirely without provenance, it should be moved to the discussion section.

-- Bob

Yes, there are also Italic languages that have the change of *k^w to p. Greek does it too. So does Romanian (quite independently of the early Italic languages). It's a pretty unremarkable sound change, so there's no reason to suppose it must have happened only once for Umbrian (or whatever it was), Gaulish, and Brythonic. All three could have undergone the change separately from each other, and in my opinion (and dare I say the standard view), they did. Angr (t • c) 14:23, 18 May 2006 (UTC)

Hmm. It is interesting to note the dialect geography of the groups that show labial vs dorsal reflexes of the PIE labiovelars. In each case, the "velar" (conservative) bunch were at the periphery of the IE territory (Ireland, an isolated outpost in western Italy around the mouth of the Tiber, Mycenae/Paros), with the "labial" (innovative) bunch inboard (Welsh, Gaulish, Sabellian in the rest of Italy, mainland Greece). This generalization is shakiest for Greek, where the velar reflexes of the labiovelars may merely be chronological. (And remember that already in Proto-Celtic, PIE *gʷ merged with the reflexes of *b and *bh.) In any case, the evidence seems overwhelming to me that Germanic, too, had P and Q dialects, given the number of good etymologies with labial reflexes of labio-velars: "four, five", as mentioned, though they might have other explanations; "wolf", "oven", etc; "sheep" < *sḱēgʷo- (= Sanskrit chāga- "goat"), "bane" (OE bana "murderer" < *gʷhen-). There's even some hesitation, as one expects in dialect borrowing: Old Norse ylgr "she-wolf" < *wḷkʷ-íH- (? = Skt. vṛkīs) (Verner's Law and all!). Certainly, if labialization of labiovelars was something "going around" in Western Europe, like the flu, there would be no reason why a group of Very Early Germans might not have gone with the flow. And nothing to say that the results would have to be attested in their entirety, by which I mean as a separate branch of Germanic showing P-reflexes exclusively or predominantly. For all that, the agreement from branch to branch is good, though actually some variation as in the "oven" words (labial in WGmc, velar in East and for the most part in North) is actually welcome.

The Romanian business has little enough to do with the case, apart from indicating that an interchange of velars and labials (in that direction only) is nothing rare. It's a sound law affecting dorsal + apical clusters, both stops (Romance *factu "act" > Rom. fapt) and nasals (Romance *leŋnu "wood" > Rom. lemn). (Note: I am unaware of any change of the type [p] > [k]; what might look like such a thing, Old Irish secht [nasalizing] "seven" < *septṃ is more likely to continue something like *sehtan.) Alsihler 21:59, 5 February 2007 (UTC)

Kündü Ghur proto germaic?

-The ugor(hungarian) "Kündü" =(Sacral king) and the ujgur Ghur = "goverment" What your opinion abuot your examle of king or kingship. /kunin~gas "king", which closely resembles the reconstructed Proto-Germanic *kunin~gaz./

I think so the recunstruction lead us to fals resolve. The realy same words usualy later generated then the languange is mede and born. The side-by-side dwellers can change the words but the ugor:kündü <-> nordic:kuningas may later exchange.

Please try use another word pl originál life stile fish~ing sail~ing hunt~ing, hut, or meat-miaso-muscle-()hús.

I don't wish to be unkind, but could I persuade you to take your question to someone who knows English better than you do, and get them to translate this for you from your own language (which I am guessing is Hungarian)? You are very difficult to understand. I think you are pointing out similarities between Germanic words and those with similar meanings in other languages. Some such similarities are coincidences, many others are words borrowed from one language by the speakers of another. Copey 2 13:20, 31 August 2006 (UTC)

I think it's probably just a coincidence. English "house" and Hungarian "ház" look remarkably similar, but they are unrelated - the "h" in "ház" is the result of consonantal shift from Finnic "k" (Finnish "kota" = hut). Compare: Hungarian "három", Finnish "kolme" ("three"). This only applied to back-vowels however; for front-vowels, Hungarian and Finnish retain the "k". Pobbie Rarr 02:39, 1 September 2006 (UTC)

Definitely a chance resemblance (which abound, by the way; I have the impression that not even my colleagues and worshipful antecedents were as conscious of this fact as they should be). The test is that *kuningaz has flawless etymological foundations entirely within the history and structure of Germanic: *kuni- "people, race" (from PIE *ǵṇH₁- ... a root than which there is no whicher, plus the possessive suffix *-inga-). Part of the point of seeing Proto-Finnic *kuninkas as a borrowing is not merely the similarity to the reconstructed Germanic form, but the fact that that it HAS no structure in Finnish, any more than sassafrass has structure in English (and for the same reason). (By the way, the problem with Hu. ház is not the initial consonant, which (continuing PFiHu. *k fits just fine with the Germanic scene) but the -z- < *-t-, which cannot be reconciled with the Germanic "house" words.) Alsihler 22:11, 5 February 2007 (UTC)

The Voiced Velar Nasal

I thought this was allophonic, not phonemic, so why is it listed as a phoneme? Bryan 11:23, 10 August 2006 (UTC)

Fixed. User:Angr 11:27, 10 August 2006 (UTC)

Intervocalic hardening

In the article,

"On the other hand, intervocalic hardening is the rule in High German (NHG habicht < OHG habuh : NE hawk < OE heafoc), and has also played a role in the later history of some of the Scandinavian languages (Sw. fjäder < OSw. fjædher : NE feather)."

In reality, it was the opposite which happened in Swedish. The Old Norse ð [ð] did not get hardened to d [d] in Swedish, but softened to [ ], i.e., it became silent. But in the orthography one kept the d and after the softening had occured in everyday spoken Swedish, one started to read the d in written material as [d]. So, the hardening in Swedish is actually due to an archaic orthography rather than actual development in spoken language. The hardening is "artificial" and what actually happened in Swedish was a softening! (NB: The hardening did occur in a few dialects, especially the notable Uppland dialect on which Modern Swedish is mainly based. One dialect, Jamtlandic, has preserved the [ð] intervocalically in short stemmed words, and one dialect, Älvdalsmål, has more or less consequently preserved [ð] in all situations.) Jens Persson (130.242.128.85 20:39, 23 October 2006 (UTC))

Centum - Satem connection

The last line in the "Hybridization as conjectured cause" section is now my own, and I can reference it (the wording is my own, as I have summerized it and not wanting to cover what that section already does). I noticed that section is asking for it to be sourced … most of it pretty much checks out by me. Nonprof. Frinkus 09:38, 5 November 2006 (UTC)

Recent change about origin

The source actually said 5, 2 distinct branches of western and northern Germanic had formed by that point (a total of 5 groups of Germanic people at this point). In addition, it also stated the evidence pointed to settled points all along the south end of the Baltic spanning from the Netherlands to the Vista at 750 BCE. It did not say there was a migration bringing them there ending at 750 BCE. Technically how it is currently written no longer jives with the source I sited. Could this be updated again, fix that? Nonprof. Frinkus 03:33, 22 November 2006 (UTC)

St <> F

I don’t understand this auditor. His contribution is a list of reverts. I may understand reverts in article, but why in inquiry for discussion? Nasz 04:45, 9 December 2006 (UTC)

Protolanguages are hypothetical and reconstructed

In linguistics the term proto language is used about a hypothesized common ancestor of a group of languages that are assumed to be related - the only way to know anything about a protolanguage is to reconstruct it using the comparative method - however such a reconstruction further adds to the hypothetical quality of the protolanguage because reconstructed forms are no representations of actual language but simply qualified guesswork at how such a language ma have been. The idea that germanic languages are descended forma common ancestro which we call Proto-germanic is a hypothesis but it is a very well founded hypothesis, and a hypothesis in which almost everyone believes. However advances in the reconstruction of the language happen each day, and there are differeing views as to it's reconstruction. This makes Proto-germanic a hypothetical and reconstructed language. Maunus 15:42, 11 December 2006 (UTC)

Every statement in science is a hypothesis, at least if you ask Karl Popper. Would you call the last common ancestor of you and any relative of yours hypothetical? The word is clearly redundant. You could add it to practically every science related article definition and be technically right. But would it make the encyclopedia better? My choice - "assumed" - was taken from another encyclopedia. I spend a better part of an hour looking throught proto-language articles of two encyclopedias but I never found any above mentioned adjectives used except for "assumed". I will not revert you but I cannot say I like that choice. Friendly Neighbour 15:59, 11 December 2006 (UTC)

I think in the articles on proto-languages that it is important to make it clear that protolanguages are not actual languages in the sense as the language we are speaking or reading. Sometimes people make claims of protolanguages having such and such properties or attempt to decipher undeciphered scripts as if they were written in a "protolanguage" - this they only do because they fail to understand the ramifications of protolanguages necesarrily being "constructs" based on the comparative method. If we were to achieve actual knowledge about the language the germanic people spoke when they were unified (if they were ever unified), say we find a way of playing their language from the soundbits recorded in potsherds, then that language wouldn't be proto-germanic but it would be a newly discovered germanic language that would allow us to reconstruct proto-germanic further back in time and with greater chances of succes. Any how I understand your viewpoint, because it is heavy words to put in the lead paragraph of the sentence and may confuse rather than enlighten. I will think of a way to better it. Thanks for your input.Maunus 16:27, 11 December 2006 (UTC)

Good job. I like your new version. In the meantime I did my homework reading a paper on the subject (Isodore Dyen, 1969 "Reconstruction, the comparative method and the proto-language uniformity assumption", Language, 45, 3, 499-518). I learnt that we both mixed up two distinct entities: occurent proto-language (OPL) as Dyen calls it (the actual spoken language being the ancestor of the relevant family) and its reconstruction, reconstructed proto-language (RPL) (which may exist in many forms). Only the former may be hypothetical (though I am happy you chose assumed in your latest edit). Only the latter is reconstructed. It seems I was right when I had the intuitive feeling that hypothetical and reconstructed cannot be right at the same time. Generally, I believe that while editing I thought more about OPL and you about RPL but we both were a little confused about the two.

I think I learned today something new. The time used for researching this was not wasted. Thanks for making me do my homework ;-) Friendly Neighbour 18:54, 11 December 2006 (UTC)

Glory be. If you can make any sense out of that Dyen article (or any sense worth taking the trouble over), you're a better man than I am. Look. We have some laboratory tests for the comparative method. Not many, but a few. The reconstruction of Proto-Romance is instructive both regarding how close it comes to what we legitimately surmise about the ancestor of the Romance Languages from other sources, and how much it misses Latin. There's a "three-body problem" here, of course, since surely no one thinks that Ciceronean Latin is the immediate ancestor of Proto-Romance, though it's disappointing that the actual Proto-Romance vowel system (i.e., much the same as Latin's) cannot be recovered by the comparative method, which pulls up only nine structure points and no hint of the feature of length. But there are some minor triumphs. Our understanding of Latin sound laws makes Latin hodiē "today" a bit of a problem, since it "should be" *hoiiē. Well, imagine our gratification when Italian oggi "today" must reflect PRom. *hoyye and can't reflect *hodye. Similarly, our understanding of Latin sound laws would predict *norus as the form for "daughter-in-law", not nurus as actually occurs. But lo: Proto-Romance "daughter-in-law" as reconstructed by the comparative method is *nora. Perhaps the most famous of such things is Shleicher's observation that the 1sg sg. thematic optative of Greek, -oimi, is probably a replacement of more original -oia. You will notice that I didn't put a star for -oia, and that's because such a form is in fact actually attested. (This -oia is itself a replacement of a still-earlier form, as it happens.)

Anyhow, yes, of course, with reference to both the above and the following, the details of a proto-language ARE hypothetical and based on evidence and reasoning that are all subject to change, and which will be argued over because of indeterminacies of various kinds (such as how many laryngeals PIE had, or how to integrate new discoveries into existing systems, such as happened with the discovery of Anatolian languages). Personally, I do not see that, once one has made clear that one is dealing with hypotheses (i.e., the products of reasoning and principles) in the comparative method, that one needs any term other than "reconstructed"; terms like "assumed, suppositional" and so on are at best elegant variations that only seem to add subtlety or legitimacy to the debate. Alsihler 22:36, 5 February 2007 (UTC)

Re: today's edit by User:Nasz: If assumed is not a good word in the context of proto-languages, please tell it not only to Encarta Encyclopedia (from which I borrowed it) but also to Prof. Wolfgang Wildgen from Bremen University. In his lecture Basic laws of linguistic and visual semiotics: From the Paleolithic to Picasso he has a slide titled "A manual dynamic archetype of an assumed protolanguage". On the other hand, hypothetical has as one of its meaning suppositional, uncertain which could be stongly misleading in the context for a reader of none or poor scientifical training (that is most of us human beings). Friendly Neighbour 15:12, 12 December 2006 (UTC)

I am inclined to agree that assumed is an alright word because it is easier to understand for the casual reader than "hypothetical recnstructed", and at the same time makes it clear that this is not an actually occuring language like english or piglatin. I have added hypohetical and reconstructed to the following sentence though which I think serves to underline the actual scientific status of a "proto-language"Maunus 15:35, 12 December 2006 (UTC)

As I've already stated above, I do agree that any proto language (in the RPL meaning) is reconstructed. I also agree that what the RPL tries to be reconstruction of is a hypothetical OPL (a language actually spoken by a group of humans at some past time). However in the case of well researched language groups - like the Germanic languages, the word hypothetical should not be too prominent as I do not think there are any modern language scientifist who would deny the Germanic languages are a family. And as long as they are a language family, they must have had a common ancestor. Full stop. On the other hand, I do not see any way the word hypothetical can be used for the RPL. Saying that someone advanced research on a hypothetical reconstructed language is like saying (s)he wrote a hypothetical paper or composed a hypothetical symphony (they all are only creations of the human mind, after all). Therefore saying something is a reconstructed hypothetical language is one mental shortcut too far for me. Friendly Neighbour 16:14, 12 December 2006 (UTC)

Interesting discussion: I wish I had time to lounge here longer. On the topic of reconstructed hypothetical proto-languages, "assumed" or otherwise, I just want to point out one thing: language communities rarely (if ever) speak a 100% totally uniform language, no matter how small or how conservative they may be. There are always variations between individuals and subgroups. Look at "English" now. Nobody would argue that the hundreds of millions of modern English-speakers all use precisely the same language. 2500 years from now, some enterprising linguist will try to "reconstruct" 21st-century English on the basis of the descendants of the language, using tried-and-true principles of reconstruction. Will that person be reconstructing a rich, non-homogeneous tapestry of overlapping dialects? Look at the comparisons of actual, attested Latin with the reconstructed "proto-Romance" language. Friendly Neighbour uses the term "common ancestor" and immediately follows it with "Full stop." I think you are missing an important point. Cbdorsett 05:20, 8 February 2007 (UTC)

Timing of the First Sound Shift

The article states that the first sound shift must have been complete by the 2nd Century B.C. at the latest. The Celtic loan words provide a TPQ. What provides the TAQ? Jacob Haller 22:40, 7 February 2007 (UTC)

Don't Latin authors as late as Tacitus often use C-forms and Ch-forms when rendering German words, tribal names, etc.? e.g. Chatti for the Hessians? Jacob Haller 22:40, 7 February 2007 (UTC)

What would qualify as a TAQ? presumably early and datable borrowings from unshifted Germanic into Latin or whatever. Or early inscriptions likewise showing no shifted consonants. There are obvious problems here, since it's hard to identify items as specifically Germanic without the consonant shifts. And the earlier the attestation, the less "characteristic" it is in all regards. "Picenian", formerly thought to be an Italic language, looks a lot like Celtic before several specifically Celtic innovations, but who knows.

Truth to tell, much about this whole entry, the bits that deal with historical linguistics specifically, are painfully confused. That may be in part the consequence of editing and tinkering from various hands of varying competence in the field. Much about Germanic historical linguistics is problematical, much (gratifyingly) is quite clear. But it is jarring to encounter such a statement, whatever the source, as the claim that Germanic was "the last" to branch off from Proto-Indo-European. This is meaningless. It has nothing to do with the way speech communities diversify. It's quite another thing to assert that Germanic is quite strikingly conservative, but it shows many and deep novelties as well. (None of which, I rush to add, provide the least support for the "creole" hypothesis.) In any case, the many slightly wonky statements about PGermanic morphology and phonology, and many of the details themselves as presented in tabular form, will not set scholarship back, nor would straightening out such wonkiness promote knowledge, truth, and happiness. So I don't know; it would take a month at least to clean the stables, and then of course folks would start fiddling. It's the name of the game in Wikiworld, but it discourages hard work on the part of people who (for example) know better than to quote Hans Krahe as an authority. Alsihler 02:43, 8 February 2007 (UTC)

It definitely occured before the 4th Century A.D. when we have extensive Gothic literature using a modified Greek alphabet (with special characters for h, hw, th, etc.). It presumably occured before the 3rd Century A.D.. Until the 3rd Century A.D., the Germanic languages were relatively closely spaced in north-central Europe, and innovations could spread from one Germanic language to another. After the 3rd Century A.D., one Germanic language had spread out to the east, so Central-European innovations would have trouble reaching it, and Eastern-European innovations would have trouble reaching the others. (However, it's possible that other Germanic languages had spread southeast before Gothic and before the 3rd Century A.D.). Aside from that 'sometime before the 3rd Century A.D.' what do we have? We're still five centuries after the claimed last possible date. Can we date the Latin borrowings? Might the same Latin contacts which bring these words into Common Germanic stabilize these words against sound-shifts? I know I don't know all the linguistic details but I'd like to see whether the dates have any evidence. Jacob Haller 05:20, 8 February 2007 (UTC)

i've normally heard the limiting dates as 500 bc and 100 bc. the 500 bc date is very tentative and based primarily on the word "hemp" *hanap from greek "kannabis" -- certain greek authors state that hemp was unknown in europe prior to 500 bc. 100 bc is based on words in greek and latin texts. Benwing 04:01, 19 June 2007 (UTC)

That doesn't make much sense. The word comes from Proto Indo-European *kanap, not Greek, and is shared by at least 4 IE branches. KelilanK 05:53, 19 June 2007 (UTC)

Personal pronouns

Aren't the personal pronouns reconstructed? Would be interesting to see... 惑乱 分からん 15:31, 24 February 2007 (UTC)

The Map

Why doesn't it show the Przeworsk and Oksywie cultures? Admittedly these didn't exist by 500 BCE but they did exist well before 50 BCE. Jacob Haller 18:59, 24 February 2007 (UTC)

Requested move

Notes and references vs. References

Currently this article has Proto-Germanic#Notes_and_references and Proto-Germanic#References sections.

• Notes and references seems to contain only citations. Would it be useful to rename it "Citations" or "Footnotes" or something else to more accurately describe it and avoid confusion with the "References" section directly below it? I am not aware of any Wikipedia-wide convention for naming such sections.
• Is the Reference section simply books that have to do with the topic at hand or are there actual reference to these works within the article? If the former is true then I shall commence updating it with about a dozen or so works that are fairly standard in Germanic philology. If not then my first question becomes even more important.

Thank you. Trollaxor 13:22, 20 September 2007 (UTC)

Unless someone objects, I am changing Proto-Germanic#Notes_and_references to Proto-Germanic#Footnotes to rid us of the redundant section names. (See WP:CITE#Section_headings for appropriate names for footnote section headings.) In the future, we should footnote the data from Proto-Germanic#References if possible and merge it with Proto-Germanic#Footnotes, thereby increasing this article's citations. The way it is now, Proto-Germanic#References only works out to be "further reading." Trollaxor 23:50, 20 September 2007 (UTC)

I will do it regardless of anyone's objections. There is no question that this needs done as this entry is a complete mess. The sooner started fixing it the better. Dr. C.S. Lewis-Barrie, Ph.D. 04:07, 21 September 2007 (UTC)

Creole Shmeole

This section is totally unreferenced and looks pretty much like the author's opinions so I am exercising my perogative to take it out. For those who are interested, here is my take on it. First of all, the section says nothing really and what it does try to say is already covered in the substrate article. So I would like to sweep it under that rug. Second, hyridization is not a serious theory and never has been. But, you have to understand, there is no hybridization theory really. Concepts of creole and hybrid are still in the process of being defined but in general they seem to mean "something" other than genetic descent (except for one fellow who claimed Proto-Germanic is a creole but that did not affect its genetic status. Oh yeh?). So, exactly what the theory is supposed to be is vague and undefined. "Substrate" is a serious theory but that is in another subsection which does not need this one. The most serious meddling of creolists (whatever that means) in Indo-European linguistics I saw on the Internet was a certain lady claiming middle English is a "creole" either of Danish or of French. Since that is closer to home we can understand it better. Well, which is it, Danish or French? Why not Latin? The whole concept seem to be trying to relate what has happened to European languages in some ex-colonial populations to the evolution of Indo-European languages. Missing are the universal principles and definitions that guide classical linguistics. Now, if we are going to develop the concept of hybridization on Wikipedia then it does not belong here which is not for innovation. Gee, it took more space to say all this than the section takes up. Somebody archive it and get rid of it.Dave 17:23, 6 November 2007 (UTC)

Hybridization theory

This section does not cite any references or sources. (November 2006) Please help improve this section by adding citations to reliable sources. Unverifiable material may be challenged and removed.

Some^[who?] also suggest that Proto-Germanic may have arisen somewhat as a Creole language due to cultural diffusion among geographically static indigenous population groups. However, creole languages ordinarily do not reflect the inflected character and the homogeneous forms of the Germanic languages.

It has also been suggested that proto-Germanic arose as a hybrid of two Indo-European dialects, one each of Centum and Satem types though they would have been mutually intelligible at the time of hybridization.^[who?] This hypothesis may help to explain the difficulty of finding the right place for Germanic within the Indo-European family. However, the Germanic languages are commonly classified as Centum languages, because of the words *hund, not **sund ("hundred", ~ centum with guttural fricative according to Grimm's law) and *hwis, not **his ("who", ~ Latin quis). That is, Ancient PIE *ģ and *ģh became PIE *g and *gh and then Proto-Germanic *k and *g instead of being turned into palatal sounds.

A relationship has been shown between proto-Germanic Centum and proto-Balto-Slavic Satem due to their close geographic proximity when they split from their Indo-European source.^{[citation needed]}

Some excisions from the Wikipedia Proto-Germanic article

Here is an excision:

• The Proto-Germanic consonants /b/, /d/, /ɡ/ are often said to have "originally" been fricatives and later to have "hardened" in some places into stops. This is disputed, however, by those who assert the opposite.
  • The main theoretical argument in favor of the "originally soft" theory is that Verner's law works out slightly neater – voicing applied to unvoiced fricatives produces voiced fricatives, which merge immediately with existing voiced fricatives. With the "originally hard" theory, the newly voiced fricatives would not be the same as the original voiced stops, and therefore a subsequent step is required to merge them.
  • The main theoretical argument in favor of the "originally hard" theory is that intervocalic "hardening" of voiced fricatives to stops is rather less common typologically than softening/weakening of voiced stops to fricatives; the most common change to intervocalic voiced fricatives is not hardening but further weakening, to approximates or to outright deletion. (Cf. common pronunciation [en to lao] of Spanish en todo lado [en toðo laðo].) Indeed, the later history of voiced fricatives in the Germanic languages often does show intervocalic weakening (OE /ɣ/ > /w/ or /j/; OE /v/ lost in hēafod > NE head, hlaford > NE lord). On the other hand, intervocalic hardening is the rule in High German (NHG habicht < OHG habuh : NE hawk < OE heafoc), and has also played a role in the later history of some of the Scandinavian languages (Sw. fjäder < OSw. fjædher : NE feather).

Reasons are as follows. No refs of any kind at any location either in notes or mentioned in the text. In addition to that the text is replete with weasel words. "This is disputed by those who assert the opposite" is a paradigm. Is there any other kind of dispute? After not giving any credible theses to dispute the author does give us the credible replies, waxing sententious: "Indeed the later history of voiced fricatives ... often does show intervocalic weakening ..." How's he know? Whoever said it didn't? What has that got to do with hard and soft? Enough of that. On to the theory. It appears to me that the application of the terms hard and soft to this instance is original research. I don't find any authors at all using them in that sense. I did find one author who considered aspirated soft and unaspirated hard, but that is not the same application at all. In all fairness, I do understand what the author means by them. He seems to mean the same as the English hard and soft consonants, but this terminology is not applied to this case of the Grimm shift by anyone I could find. If you can find one then perhaps we can work it back in. And now we read of the "main theoretical argument in favor of the originally soft", etc. You would think a main argument would appear in most books on the subject but not this one; in fact, I can't find anyone who uses any of these arguments at all! And my book list is getting up there, as you can see in the notes. Well, as far as I can tell, there isn't really a dispute, as everyone admits the situation is somewhat ambiguous and moreover those are not the only possibilities. So I am just going to replace this with a quote or two from one or more of the linguists' previewable books. There are a few general references below but - gosh by golly - the author has chosen only non-previewable books so we'd have to pay through the nose to take a look. I do have have one of his books, Bennett on beginning Gothic, but I'm going to take out that Gothic stuff too because this is on Proto-Germanic not Gothic.Dave 23:11, 10 November 2007 (UTC)

Another excision:

• • Evidence from all branches of Germanic shows that /ɡ/ was [ɣ] elsewhere, including initially. Initially it was "hardened" to [g] independently and at various times in the various languages:
    • Before 350 AD in Gothic (early borrowings indicate lack of initial [ɡ]).
    • Before 1000 AD in Old English (palatalization of initial /ɡ/, c. 450 AD, is consistent with [ɣ], not [ɡ]; similar arguments apply to Old Saxon and Old Frisian).
    • Perhaps as a result of the High German consonant shift, before 800 AD (Low Saxon dialects still have intervocalic [ɣ]); but some linguists have asserted that /b/, /d/, /ɡ/ have always been stops in all positions in High German.
    • Perhaps before 800 AD in Pre-Old Norse, when Old English speakers began borrowing words from Proto-Norse.
    • Not yet, in Dutch.
  • Evidence differs with regard to /d/. In the oldest representatives of all branches of Germanic it appears that /d/ was a stop [d] initially, or when geminated, or after a nasal. In Gothic and Old Norse /d/ was a fricative elsewhere, [ð] (except where it came into contact with a voiceless consonant in Old Norse, and finally in Gothic, in which case it was devoiced to /θ/). But in West Germanic /d/ became a stop [d] in all positions. Note, then, that Gothic and Old Norse show a symmetrical system where /b/, /d/, /ɡ/ are stops when initial, doubled or post-nasal, and fricatives elsewhere. The reconstructed system of the other (West Germanic) dialects, however, is highly asymmetric (/ɡ/ is mostly fricative, /b/ is part stop, part fricative, and /d/ is entirely stop). Analogy works towards symmetry, and hence the reconstructed West Germanic system is likely to be correct and the symmetric systems of Gothic and Old Norse secondary developments. (An additional argument for this is that early borrowings into Gothic corroborate the initial [ɣ] in Pre-Gothic as in West Germanic.)

These changes do not concern Proto-Germanic, the topic, but are the history of those sounds through all the periods. Some interested editor should make sure they are in the articles for those languages.Dave 01:13, 11 November 2007 (UTC)