7
  1. First, I don't speak/understand any so-called agglutinative languages, like Turkish. I also don't know German.

  2. I understand there's no good definition for the concept of "word", which could apply to all languages. But for the sake of this questions let's assume we define word as something independent of writing.

  3. I am always a bit skeptical about some languages being able to construct very long words. This skepticism come from the fact that I've seen such claims about languages I know well. For example: many times I have heard and read info like "English words are much longer than Chinese and Vietnamese words" while in my opinion these claims are based on the biased bracketing of the syllables and morphemes. For example a word like: "unbreakable" translates to Vietnamese "không thể phá vỡ", for some reason "không thể phá vỡ" is never regarded as a word neither by Vietnamese nor by English speakers. I would not consider "unbreakable" a single word any more than "không thể phá vỡ".

Now I think similar "tricks" are made with German and Turkish long words. That is I believe that the Turkish "muvaffakiyetsizleştiricileştiriveremeyebileceklerimizdenmişsinizcesine" and the German "Kraftfahrzeug-Haftpflichtversicherung" are just bunches of words. It just happens that spaces are not put inside these bunches.

The only doubt I have about my example is that "không" is actually a single word that can be uttered alone, while the English prefix "un" is not.

Could anyone break down these long words to shorter words to show it's possible or contradict my intuition-based claims otherwise?

25
  • 28
    I don't think you'd find many/any native English speakers who wouldn't consider "unbreakable" to be a single word.
    – Cairnarvon
    Jan 15 at 23:16
  • 11
    Un- is a prefix and -able is a suffix. One of the best definitions of ‘word’ that I’ve come across is that it can be meaningfully uttered in isolation and used to fill a slot in the syntax of the language in question. That’s true of break, but not un- or -able (which is different from able). In the German example, it works fine for Kraft, Fahrzeug, Pflicht, etc.; those are all words, and the whole thing is a compound. But why shouldn’t (German) compounds be words? They can fill a syntactic slot, and cases are added at the end of the whole thing, not per element. Jan 15 at 23:27
  • 6
    English itself has many single "words" that could be analyzed as two separate nouns: "innkeeper", "sailboat", "football", "bookshelf", etc
    – chepner
    Jan 16 at 13:59
  • 6
    That dash in Kraftfahrzeughaftpflichtversicherung just hurts. It doesn't belong there. Yes, with the most recent spelling rules it's allowed to add dashes for 'better readability' (by people who don't read well, if you ask me), but still.
    – Nobody
    Jan 16 at 18:53
  • 6
    A german native speaker would consider Kraftfahrzeughaftpflichtversicherung to be a single word. It doesn't really matter if you don't think so. Also, scrabble is played with words that can be found in a dictionary. Kraftfahrzeughaftpflichtversicherung can be found in a german dictionary.
    – Polygnome
    Jan 17 at 1:12
40

From the perspective of linguistics, the question is meaningless though well-intentioned. "Word" is not a well-defined technical concept in linguistics (or, some people may have concocted a definition of "word" for their purposes, but there isn't even a widely-believed definition). The best definition is "a maximal string of letters not containing spaces", and that's really not very good (not all writing systems use spaces, plus that means that unwritten languages don't have words).

One approach has been to equate "word" with "syntactic terminal", which simply swaps the question to "what is a syntactic terminal?". In contemporary Minimalist syntax, verbs especially are composed of many many nodes which still are realized as "a word" in a language like Latin (with tense, aspect, person and number agreements each of which contributes a syntactic node). There are also phonological accounts which appeal (circularly) to some property of assumed words (stress on the penultimate syllable of the word; a requirement to have at least 2 syllables in the word).

It seems to be true that anything that a naive speaker of a language can utter alone is at least a word, therefore an English speaker can't say "gira" which is a sub-part of "giraffe" and a Saami speaker can't say "beatna" which is a sub-part of [beatnag-a] "dog (acc. sg)". Clearly, not every utterance is a word. If you define "word" as a minimal utterance, then you would exclude the German and Turkish examples as "not words" (they are not minimal). But then "oxen" and "cats" are not words, because then contain words – "ox, cat". And that just seems wrong.

There does appear to be a phonological object, a grouping of many syllables into a thing, where rules apply within that thing, or with reference to the thing ("no obstruents at the end of the ___"), which we call the phonological word, or ω. This thing isn't "defined", it is or may be constructed, with language-specific rules. When you find that a large sub-part of an "utterance" has a certain kind of coherence w.r.t. phonological rules, you can call that unit a "(P-)word". But it turns out that this ω thing is not always coherent in a language, and clitics can present contradictory evidence where they partially act like they are "in the word" and partially act like they are outside.

Hence, most linguists have abandoned the concept of "word" as a coherent technical concept.

7
  • 6
    This is a good answer, except for the puzzling claim that the definition of a word as a string of letters with no spaces is the best one. There’s no satisfactory definition, agreed, but that one has to be one of the worst ones: most have more or less peripheral problems that depend on more or less arbitrary categorisations of different aspects of language, but this one is perhaps the most arbitrary of all and excludes the vast majority of all words in its very definition. Jan 16 at 1:55
  • 13
    But that is how most people understand "word", so if you ask a random person "how many words in this test", they will count blank-delimited sequences (if it's not an Arabic / Chines / Thai speaker). It correlates best to the procedure that ordinary people use for identifying words. I'm not at all suggesting that "word" is a valid technical concept, so my metric of goodness relates to understanding this strange social object that people talk about.
    – user6726
    Jan 16 at 2:05
  • 6
    True, most people will use it as a basic beginning point in identifying words if faced with a written text (until you ask them why classmate is a word and soul mate isn’t); but even the most average of random speakers would be readily able to recognise words as spoken units, which this definition excludes completely. Nobody would say you have to write it down for it to become a word. Jan 16 at 2:14
  • 1
    @JanusBahsJacquet Yet you do have to write it down mentally to answer the question whether soul mate/soulmate is one or two words in the common understanding of the concept 'word'. It sounds perfectly sensible to define this common understanding of the concept as a factor of spelling as the German language beautifully shows. The observation that languages without a writing system and illiterate people would have a less strict/non-existent understanding sounds likely/reasonable. Jan 16 at 23:41
  • 5
    @Polygnome Whether a string of letters/symbols is considered a valid word in the game of Scrabble is determined by whether said string is included in some list agreed upon by the participating players. (Commonly used word list references for the English language are "NASPA", "Collins", etc.) However, that doesn't necessarily mean words are well-defined as a linguistic or technical concept. It might just as well mean that some "authority figure" by the name of "Collin" tossed a coin every time he had to decide whether a queried string should be deemed worthy of inclusion in his list.
    – Will
    Jan 17 at 3:33
15

In German, noun phrases that are used to describe a separate entity other than their individual nouns are written without spaces. Thus, the example of Kraftfahrzeug-Haftpflichtversicherung may indeed be considered as a "bunch of words" in the sense you have described. In Turkish however, this is not true. The example in Turkish you have provided contains only a single "word" which is muvaffakiyet, meaning success. All of the remaining parts of that word are suffixes. They are used to modify the meaning, add the notion of time etc.

For example:

  • Muvaffakiyet = Success
  • Muvaffakiyetsiz = Unsuccessful
  • Muvaffakiyetsizleş (-mek) = to become unsuccessful and so on.

If you check the last part, you might think that the last added part "-leş" means "to become". But that would be incorrect. It is just a suffix and is meaningless on its own. So your example in Turkish is not a "bunch of words" written without spaces since except for the first one, muvaffakiyet, none of the components have a meaning on its own.

On the other hand, your German example, Kraftfahrzeug-Haftpflichtversicherung, translated into Turkish would be "motorlu araç sorumluluk sigortası". As you can see, that is written with spaces as each of the individual parts are not suffixes, but nouns, meaning that they have a meaning on their own. (Motor (-lu) = (with/containing) motor, araç = vehicle, sorumluluk = liability, sigorta (-sı) = insurance (of))

6
  • 2
    Your example is continued here at the bottom: en.wikipedia.org/wiki/Longest_word_in_Turkish
    – aytunch
    Jan 16 at 19:31
  • 1
    @talkanat. Sue, leş can't stand alone in Turkish, but so are a and the in English and các, những in Vietnamese and and in Chinese. but somehow all these 6 morphemes are never agglutinated into words in English, Vietnamese and Chinese, thus making the words of these languages shorter. Can we imagine a different writing convention for Turkish words? Is the problem with the number of suffixes in Turkish?
    – GA1
    Jan 17 at 13:12
  • @GA1, as a native speaker of Turkish, I can say that, writing those suffixes separated from the stem would be writing the English word unsuccessful as un success ful. It would be impractical to separate those suffixes of the words by spaces, perhaps because of the number of suffixes in Turkish as you have pointed out.. Also, note that, the vowel of suffixes are modified in most Turkish words. For example, suffix -leş in your example would be -laş for words ending in, for example the letter ı. Example: Başarı-sız-laş (-mak).
    – talkanat
    Jan 17 at 20:02
  • Also note that "muvaffakiyet" is not a "word" either but made by adding the Arabic prefix "mu" and suffix "iyet" to the Arabic word "vafk/vefk". Ironically all those additions make the word longer but the final meaning is the same as the root word (muvaffakiyet = vafk).
    – Selcuk
    Jan 18 at 5:43
  • @Selcuk, there are no prefixes in Turkish language. And any loan word from another language that is used in Turkish is considered a word as a whole with its suffix/prefixes. That is why I denoted muvaffakiyet as the "stem" here. But considering its Arabic root, you are of course correct.
    – talkanat
    Jan 18 at 17:01
4

In German these compound terms are being called "Komposita" (which is the plural form of "Kompositum").. As per your example "Kraftfahrzeug-Haftpflichtversicherung", this could as well be expressed as a "Haftpflichtversicherung für ein Kraftfahrzeug". In order to merge such even more lengthy descriptions into a single term (without inventing a new one), compound terms are often being formed. The term "Kraftfahrzeug-Haftpflichtversicherung" can be split up into 5 words:
"Kraft" (power), "Fahrzeug" (vehicle), "Haft(ung)" (liability), "Pflicht" (duty), "Versicherung" (insurance).

The term "Donaudampfschifffahrtsgesellschaft" is probably still the most promeniert, due to the extraordinary tripple f (which only may occur in such a compound term, but nowhere else).

6
  • 2
    This does not answer the question (i.e. whether it "really" is a single word).
    – Norrius
    Jan 16 at 19:47
  • 12
    If you check the discussion here carefully, you'll see that something being splittable into multiple words does not necessarily make the compound "not a word". English has many such examples, like butterfly or classmate, which by your logic should be "obviously" not considered words.
    – Norrius
    Jan 16 at 20:17
  • 2
    The English term for a Kompositum is ‘compound (word)’. Jan 16 at 22:16
  • 2
    Technically, Kraftfahrzeug-Haftpflichtversicherung can be split up into 6 words, because the word Fahrzeug can be split up into Fahr and Zeug ;) Jan 18 at 0:38
  • Donaudampfschifffahrtsgesellschaft was already a very prominent example (most of the time used with an addidional Kapitän at the end) long before the orthography reform of 1996 introduced the third f.
    – Bill Tür
    Jan 18 at 9:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.