About this FAQ
A selection of my answers to questions on linguistics and language forums. Some have been edited & expanded for general usefulness, precision or to respect anonymity.
Answers & Observations
Are some languages superior to others in their ability to express advanced concepts or in the way they form new words?
Consider how a few languages express a fairly modern concept: the word for "computer".
French speakers coined a new word based on learned (but not necessarily "borrowed") Latin roots: ordin- "order" + participial -at- + derivational suffix -eur "-er"/"-or".
ηλεκτρονικός υπολογιστής (ilektronikós ipoloyistís)
Greek speakers coined a phrase "electronic calculator" and use it with the meaning "computer".
Turkish speakers combine the roots for knowledge & count(ing) to derive a new, native word designating "computer".
Zulu speakers borrowed the English word "computer", nativized the pronunciation & even adopted it grammatically into one of the Zulu noun classes. This is akin to the way English speakers borrow & nativize most science & technology terms from Greek and Latin.
English speakers coined a new word from borrowed Latin comput- "to count" + derivational suffix -er.
Which of these ways is "superior"? Which language is more suited to develop a word for computer? To become linguistically minded requires that you move away from what a language should do and consider more deeply what languages actually do.
Once the concept is introduced, concerned speakers in all of these languages have a way of talking about it. We have to be more specific than superior/inferior to analyze these techniques rigorously and compare them to one another. Does the language use analytic methods for coining new words ("the thing that counts" or "box of information") or synthetic ("countingthing", "informationbox", "infocounter")? Does the language tend to borrow loanwords for new concepts or coin new words for them?
I've heard of the apparent parallels between Irish and Semitic languages. Can you elaborate on some of these similarities?
Here are a couple of relevant parallels I'm familiar with:
- VSO is the basic word order of Irish and Classical Arabic.
- Lack of indefinite article, presence of definite article in both Semitic and Celtic.
- Conjugated prepositions. Arabic `alaa 'upon': `aleyya (upon-me), `aleyhi (upon-him), `aleykum (upon-you), etc. Irish le 'with': liom (with-me), leat (with-you), leis (with-him), leo (with-them), etc.
- Syntax of the genitive construct: noun(s) - genitive article - genitive noun. Additionally, both families only allow 1 definite article in a genitive noun phrase, which specifies the genitive noun. Compare Hebrew beit ha-sefer (house the-book) 'schoolhouse' and Irish Tír na nÓg (land the-Youth) 'Land of Youth'
Looking at this from the perspective of historical linguistics, these similarities likely developed by chance. Another possibility is that they reflect common areal phenomena, which requires uncovering how and when the two families had the kind of contact that could produce this result. Celtic and Semitic linguists are skeptical of the few hypotheses elaborated so far.
Some issues with linking Celtic & Semitic: the Celtic features above are unique to Insular Celtic (particularly Irish); these supposedly Semitic features don't show up in other languages more likely to be influenced (in, say, Italic by, perhaps, Phoenician); there is strong, ancient textual evidence supporting the idea that these features developed naturally in Insular Celtic; there is even stronger evidence that Proto-Semitic demonstrated most (all?) of these features prior to the existence of Celtic as a distinct Indo-European branch.
A more plausible link between the families is suggested by what possibly amount to Semitic loan words in our remote ancestral tongue, Proto-Indo-European (like PIE *woynos 'wine' from Semitic *waynu). See http://paleoglot.blogspot.com/2008/08/list-of-possible-proto-semitic-loanword.html for a sample list. It's a curious thought, since Proto-Indo-European and Proto-Semitic were likely spoken around the same time.
How do experts or linguists go about learning a language? Is their way different or better?
As with all people, "linguists" and "experts" have varying approaches and different interests that drive their individual learning goals. However, there are some key differences between seasoned learners and first-timers. In general, linguists don't search for the right program or method to giftwrap the language for them. Instead, linguists tend to focus on the language itself. This is a more active approach - talking to native speakers, listening to music, reading books, watching T.V., etc. These experts also seek out consistently sporadic exposure to the language. This means maintaining prolonged contact with the language, during which time the examples you hear are spontaneous and native, not forced. This is in opposition to the pre-planned lessons with scripted conversations that many learners just repeat for the sum total of a couple hours a week.
A popular approach is to start modeling the grammar of the language based on spoken or written data. Grammar includes pronunciation, word structure, sentence structure and use (literature, slang, etc.). Building a sophisticated, accurate grammar requires you to make A) observations, B) abstractions and C) tests.
For example, assume that you're studying Modern Greek (I use transliterated characters below rather than the Greek alphabet) -
This is the empirical part of your study, where you collect evidence for how the language works. Let's say you come across two sentences:
(1) afti erh-onte xana (they come-3pl again) 'they come again'
(2) afti xana-erh-onte (they again-come-3pl) 'they come again'
You can also check with speakers to find "negative evidence", meaning things they wouldn't say:
(3) *xana afti erhonte (again they come-3pl)
And you'll note that this word order is judged as a non-native or poorly formed. You can confirm with others to see that Greek speakers also don't write or speak sentence (3), but they do write or say (1) and (2).
This is the "thinking" stage, where you use observations to form basic rules that you'll refine over time, eventually allowing you to communicate as a fluent speaker. You can view these as fluid and dynamic bottom-up rules that emerge from some level of the language, or as top-down rules that have some central control over the language. Either way, let's say you fully understand the vocabulary and pronunciation aspects of Gk sentences (1), (2), (3) above, but you're not clear about the word order. You will try to abstract a simple rule:
(4) observation: Greek allows the semantic prefix 'again' to be prefixed to the verb 'come' or to be detached, BUT detached 'again' must follow the verb
(5) abstraction: Greek allows [Prefix] either to be prefixed to verb or to come detached, BUT detached [Prefix] must follow [Verb]
(6) Rule: Either attach prefix to V or detach from V. If detached, place after V.
C) Set up test
In the last step, you came up with a rule (6) that should help produce sentences in Modern Greek. For you to check if your explanation is any good, you need to make it a specific and testable hypothesis:
(7) When any Modern Greek verb has a prefix, Greek speakers will alternate between keeping the prefix attached to the verb and separating the prefix from the verb. If they separate the prefix from the verb, they will not place the prefix before the verb.
You can't be afraid to be wrong here - if (7) is too generic, you won't learn anything in particular, because you won't be testing anything specific! If you leave out some information to make yourself look like a better speaker than you are, you'll never figure out if your understanding of this language's grammar is correct!
With hypothesis in hand, you test against more observations (back to step A). As you repeat and repeat, you'll build a better model capable of accounting for more of the language.
The data come from everywhere - books, CDs, internet, conversations, field studies and so on. There's no reason to follow any specific commercial "program" or "product" without consideration for the quality of its linguistic data - all you want is good exposure to the language itself.
Wait, I'm lost. Can you simplify your explanation of how some experts learn languages?
The linguistic approach allows the learner to be an active observer. That means the learner makes observations. These observations are the foundation learning a new language. For example, imagine you were learning English and observed the following two sentences:
"He goes to the store."
"She goes to the store."
While listening to the language you look for patterns. From the two sentences above you might guess that the pattern for word order in English is subject - verb - object (SVO). You might also guess that 3rd person singular verbs in English are created by taking the basic verb and adding an 'es' ("go" becomes "goes"). On top of that, you might guess that English verbs don't distinguish between males and females - 'goes' is 'goes' for him & her!
Then, you test the patterns that you guessed to see if they are correct or not. Take the verb pattern above. Are there places where it doesn't work?
So here's an example where adding an 'es' doesn't give you the right verb: 'sing' plus 'es' would have given me 'singes', but the evidence is that people say 'sings'! So, you would need to refine your pattern. Determine where "es" works, where "s" works and find if there are other ways of conjugating a 3rd person singular verb you haven't yet discovered.
This method for learning languages is similar to a science experiment. You observe what's happening in the language. You hypothesize (try to find a pattern). You test the pattern. You refine as necessary. Then you repeat that experiment again and again.
This method can work better than rigidly following a book or a course, but it does require that you take lots of notes (observations!) and pay close attention to what is happening in the language.