14

The problem is, nobody is quite sure how PIE was pronounced! When we talk about PIE phonemes like /*d/, we don't mean it was actually IPA [d]. We mean that "there seems to have been a phoneme, which is pronounced [d] in a lot of descendant languages". But there are also many languages which don't pronounce it [d]: Germanic, Anatolian (an extremely ...


10

If your want a user-friendly program to use on the go you can check Cmaptool from good people at the Florida Institute for Human & Machine Cognition (IHMC) it's really easy to use straight forward and you can export your data to variety of extensions. If you TeX you can use qTree package from CTAN. \Tree[.IP [.NP [.Det \textit{the} ] [.N\1 [....


9

Update: I have cleaned up and organized this list significantly, and it is now available here. I had the same question as you, and ended up throwing together some perl scripts to scrape Wikipedia's list of the 10,000 most used French words, pipe them through eSpeak to get their IPA pronunciations, then do a simple (completely non-phonetic) comparison to ...


9

אַ שפּראַך איז אַ דיאַלעקט מיט אַן אַרמיי און פֿלאָט a shprakh iz a dialekt mit an armey un flot Max Weinreich As WavesWashSands says, the earliest known published source is Weinreich's article: דער ייִוואָ און די פּראָבלעמען פֿון אונדזער צײַט Der YIVO un di problemen fun undzer tsayt1 originally presented as a speech on 5 January 1945 at the annual YIVO ...


8

I should think that there are a number of sources. For example, the University of Sheffield offers videos for every IPA sound. If the sounds of English, Spanish and German are enough for you, the University of Iowa has an animated app/web-app with videos. Edit: I think reddit.com/r/ipa is also noteworthy, as most new apps and websites about the topic come ...


8

Max Weinreich is usually said to be the person who created the quip, but according to the Wikipedia entry on him, he was actually quoting someone else.


7

There doesn't appear to be such a list, hence I made you one based on this word list: i o aç iç öç uç üç ad od öd af of öf uf ağ eğ iğ ah eh ıh oh ak ek ok al el il ol öl an en in on ön un ün ar er ör ur as es is ıs us üs aş eş iş üş at et it ot öt ut üt av ev iv ov öv ay ey oy uy az ez iz öz uz üz be de he le ne re ve ye bu hu mu su şu da ...


6

We created eztreesee so that rank beginners can try out sentences without having to install parsers, models, etc. on their own computers. The backend runs entirely off the Stanford Parser. And therefore, what you see there are Penn Treebank tags. The official documentation for PTB tags is this weird PostScript file; you might prefer this quickref.


6

The first thing to do is stop thinking about spelling, and start thinking about pronunciation. Proto-Indo-European had no spelling, and for a long time, neither did Latin. In this case, you need to understand the sound changes from PIE to Latin (for a word like "hundred", you would need to know the sound changes from PIE to proto-Germanic). Two useful books ...


6

Here are a few suggestions: There is a MOOC being offered through Coursera: https://www.coursera.org/course/humanlanguage The virtual linguistics campus has some good content and is well taught but might be difficult to navigate for a new student. It's good if you have a topic in mind: https://www.youtube.com/channel/UCaMpov1PPVXGcKYgwHjXB3g The ling ...


6

Not a single dictionary, but you could get the effect by chaining together resources. I suggest starting with Bopp's Glossarium Sanscritum which is a Sanskrit-Latin Glossary. After a few pages of that, you will probably be convinced that for etymological study, you don't want a Sanskrit-X dictionary, you probably want a dictionary of Sanskrit with ...


6

If you're interested in inflected forms, you could take data from Wiktionary's Latin section. Most verb entries have automatically-generated conjugation tables, and all of the content is CC-BY-SA. If you're interested in meanings, you can download data files for well-regarded dictionaries through the Perseus Project. Lewis and Short is my go-to resource; ...


6

Contrary to the expectations of some commentators, doctor-patient corpora are available (under some conditions, needing to sign some licence and confidentially agreement) for research. The standard entry point for a search for such corpora is the CLARIN Virtual Language Observatory and entering doctor patient in the search slit gives currently twelve results....


6

You're essentially describing the field-worker's script. In reality, such a list is an ever-growing project constructed on the basis of previous results (when e.g. you discover that you can't use the -ááka- tense in questions, and have to figure out why). The problem with constructing a universal list of actual sentences to get is that you will either omit ...


6

The Grupo de Ingeniería Lingüística, part of the Universidad Nacional Autónoma de México has a WhatsApp corpus of undergraduate students. They have a paper that introduces it: http://www.aclweb.org/anthology/W18-3501. They have their own web-based corpus management tool http://www.corpus.unam.mx/geco/ but I don't see the WhatsApp corpus on there, so you ...


5

CELEX certainly has this information (at shallow and deep levels of morphological analysis). It is not free. I wasn't able to find a free alternative in a quick search. There are also morphological analysis tools for building this database yourself, though it's an area of active research. You might start by trying the resources here (http://aclweb.org/...


5

I would look at the Algonquian languages (Meskwaki, Ojibwe, Arapaho, Blackfoot, Migmaq etc.). They display both obviation and personhood hierarchies, and have fairly good collections of published glossed texts. To start out, check out the work of Ives Goddard (the high priest of Algonquian linguistics), Richard Rhodes, Amy Dahlstrom, and John Nichols; you ...


5

If you have access to a good library you can consult von Wartburg's Französisches Etymologisches Wörterbuch in 25 volumes. There is also an abridged French translation.


5

Machine translation in general is in its infancy. Even for major languages like Mandarin Chinese and Spanish, computers have trouble with context-dependent concepts such as verb inflection and words with multiple meanings. No machine translation is reliable. All it's good for at this time is to help you get the gist of a text in another language. Google ...


5

Two resources are particularly popular. WordNet Wiktionary


5

I have to first point out that "represent" is ambiguous in linguistics: it could refer to a mental object, such as "the representation of tone in the (mental) grammar of Chinese", or it could refer to writing practices – how to notate a thing. I take it that you are asking about notation rather than the underlying mental thing. Even so, there is a difference ...


5

The best reference source is the UCLA phonetics collection, here (you will notice a lot of other interesting sound categories). I have strong reservations about using Wiki exemplars which are not produced by a person natively uttering words in languages with the sound in question. The bilabial ejective is, in my opinion, misleading in that the performer ...


5

You will likely not find a program that "converts" natural language sentences into plain Prolog code like Green(X) :- Grass(X). because natural language is way too involved to be accuarately expressed as facts and rules when taking the linguistic predications directly as logic programming predications. For example, how would you represent mass nouns (does it ...


5

Krahmalkov's Phoenician-Punic Grammar contains a high-quality (but unfortunately short) dictionary, focusing mostly on functional words and affixes rather than content words. It seems like it's meant to be used alongside a database of triconsonantal roots. EDIT: Krahmalkov also has a Phoenician-Punic Dictionary alongside the grammar, and I finally managed to ...


4

I think it is Joseph Greenberg who first used this notion (borrowed from biogeography/taxonomy) in his controversy with Malcolm Guthrie, a leading Bantuist in the 1960s, about the Bantu homeland. Guthrie believed it to have been in the center (somewhere in what is today the DR of Congo) of the huge Bantu area (roughly all Africa south of Equator save the ...


4

The answer to the headline question is very simple. None. Do what millions of successful language learners (and teachers) do, completely ignore linguistics. Just compare linguists and historians - you'll see that linguists are no better at learning other languages. Historically, advances in language learning have not generally come from advances in ...


4

Here are some resources, in no particular order: The Wikipedia article on Speech Synthesis gives a pretty good overview on the topic, although it's a bit thin in its discussion of statistical parametric (i.e., HMM-based) synthesis. Dennis Klatt's History of Speech Synthesis gives example audio clips that exemplify the evolution of rule-based formant speech ...


4

Let me share with you an experience I had in learning Russian. I was utterly incapable of pronouncing the (stressed) high central vowel transcribed y (such as in the name Gromyko). The best I could do was the English vowel in hit. (If you're from New Zealand, this is the right thing to do, it turns out.) I asked a Russian friend to listen to me pronounce ...


4

Ironically, your best bet is: https://archive.org/download/stackexchange


4

Seeing Speech is a remarkable site that has MRI, UTI and animations for most IPA sounds.


Only top voted, non community-wiki answers of a minimum length are eligible