How do we know old/ancient pronunciations?

I have heard occasionally about the Great Vowel Shift, where English vowel pronunciation changed over a period of time. I just heard it referenced on the Lexicon Valley podcast.

OK, so how do we know how people pronounced things in the 1300s?

I often wonder the same thing about ancient Egyptian. I understand how the meaning of hieroglyphs was deciphered, but not how we know what sounds they represented.

Phonetic spelling found in 1300’s English pre-textbooks? Someone had to teach the clergy and educated class how to speak proper English.

Because we can start with the assumption that people don’t put in silent letters when they first start writing.

Combinations like “kn” in “knight” and “knife” presumably originally were pronounced.

So to “gh” in a lot of words (again, “knight”).

Then, older documents reflect pronunciations. For example, journals from people like Lewis and Clark and Peter Pond have idiosyncratic spellings, which presumably show how they actually pronounced the words.

It was the invention of the printing press, just before major changes in English pronunciation, that froze spellings that subsequently no longer reflected pronunciation.

I am often amazed when I’ve seen discussions of the history of language families. Tho I am quite ignorant of such things, I would imagine some hypotheses could be formed based on the current usage of languages which split off from or were influenced by the ancient language at different times.

Poetry is also useful. If a poem has “sound” rhyming with “wound”, we don’t know exactly how those two words were pronounced, but we know they both ended in the same sound, unlike modern pronunciation. Building up a network of such correspondences can tell us a lot about pronunciation.

Okay, I’ll bite: how do we know that assumption is valid? For example, Greek has the letters Theta and Phi, which are commonly transliterated as th and ph.

As in “silly English ka-nig-hits”? :smiley:

You can also deduce things from related words in other languages. For example, the English word “knee” is obviously related to the German word “Knie”, which means the same thing and in which the K is pronounced. That’s very strong evidence that the K in the proto-Germanic word for “knee” was pronounced, and that the letter stopped being pronounced some time in the evolution from Anglo-Saxon to modern English.

Sometimes music with lyrics can help too.

e.g in This is the Record of John by Orlando Gibbons (early 17th century), you can see from the music that -ed at the end of words was usually pronounced as a separate syllable.

“And he confess-ed and deni-ed not”
“And they ask-ed him…”

This is confirmed by other lyrics by other composers of 16th and early 17th centuries. You can also see from other musical settings that ‘temptation’ was pronounced as 4 syllables, ‘temp-tat-i-on’, not ‘temp-tashun’ as we pronounce it today. There are many other examples.

Right, but those form a digraph. “Th” is pronounced as one sound, because the stock Latin alphabet didn’t have a letter that represented “Th” sounds. And in Old English they added Edh (Ð, ð) and Thorn (Þ, þ) to represent those sounds. The letters got dropped but there was still a need to represent the sounds and so sometimes Y was used, and sometimes the digraph TH.

So yes, at one time “knight” really was pronounced “kunicht”. But even there I have to use a digraph to represent the middle guttural sound. Gh, ch, or ƿ are ways to represent that sound. And now our modern pronunciation doesn’t even pronounce the fossilized gutteral, or the initial “k” sound.

But anyway, of course in old english or middle english times there was no such thing as a standard orthography. People spelled words the way they thought those words sounded, based on orthographic rules that are sometimes frankly ridiculous, like silent “e”. So we should spell the word that means guys who wear armor and swing swords “niit”. Double vowels would at least make sense. Or we could do what some languages do and put hats on every other vowel.

This part of the question is really interesting. Anybody have any info on this?

I believe this was a central plot point in the movie Stargate.

Cecil’s column:

The short answer is we started with the Rosetta Stone. Not only did it provide a way for us to translate the hieroglyphs, but it also provided a lot of clues about pronunciation. You start by looking at names and how they are spelled out or symbolized in each language.

Cecil goes into a lot more detail. It’s definitely an article worth reading.

Based on a brief skim of the Wikipedia article I’d say it’s incomplete, but not that hard, since, although it’s based on pictures, most of the writing uses those pictures to represent specific phonemes, and some things are known of Egyptian words from non-Egyptian sources and Coptic, a descendant language.

I’m more baffled at what can be done to figure out what Ancient Chinese sounded like and how we can figure it out. (An excellent YouTube video.)

You may be interested in Cecil’s answer on the Straight Dope.

How do we know how to pronounce ancient Egyptian?

That is a good video. Seems in each case though a necessary element is pronunciation in modern descendant or influenced languages. And the logic mentioned for English spelling that it’s less likely for spelling to start out with silent letters. Similarly as in that video it would be strange if Korean, Japanese and Vietnamese all had an ending ‘k’ sound for the character for country, which modern Mandarin doesn’t, but those languages all added it rather than older Chinese pronunciation being what those languages picked up*. From what I’ve read Coptic sounds are the main guide to Ancient Egyptian sounds. But as in the otherwise silly movie ‘Stargate’ mentioned, there’s no way you could know if they had remained the same without going back in time and hearing them.

Same AFAIK with Mayan and figuring out the phonetic use of the glyphs. It’s based on the similar language people of the region still speak. If a language was somehow completely wiped away you might not be able to figure out much at all.

*in Korean you can go back to the 15th century and see how Chinese characters were spelled then in the newly invented alphabet, not only how they are now. The spellings are sometimes different or vary more than now. Although OTOH it runs into the problem of not being sure how people pronounced hangul letters back then. For example the character ‘大’ (big) now has no ending vowel in Korean or Chinese. But in early hangul texts it’s sometimes spelled ‘땡’ (ddaeng) rather than ‘대’ (dae) as it’s always spelled now. But it’s not clear an ending ‘ㅇ’ had the same sound then, might have been closer to silent in that case back then than it would be now.

Thanks! I have a question about his part of that article:

How did they know that the non-hieroglyph languages on the Stone were properly equating the same sounds? For instance, if there was a German/English rosetta stone, would the researchers have determined that the Germans pronounced their country similar to the English pronunciation of “Germany”?

Stretching back even further, what theories are there about how people communicated verbally before written language? How far back before communication was essentially grunts and hand/body motions?

I’m not sure what you mean by your first question. Before written language, people talked with their mouths, like they do today.

They didn’t know. But they could hypothesise. And it’s not like they drew the conclusion from doing one name. If you did:

Italy - Italia
France - France
Germany - Deutschland
Spain - Espana
Portugal - Portugal

with the latter of the pair in some unknown script, you could figure out what was useful information, what was likely something similar, but of totally unknown pronunciation, and what was probably a completely different word.