It seems to me that there is a problem in principle with Esperanto's combinatorial structure, which is rarely explored.

Whilst it is true that the meaning of many compound words is fairly transparent from the components and a knowledge of the world, there are cases where the combinations are unsuitably vague in their application.

This will naturally arise in areas where there is insufficient usage to attach a specific meaning to the compound, and the real world offers several distinct (and to be distinguished) applications of the compound.

Ploughing through an English-Esperanto dictionary, such as Wells or Benson, will reveal the number of occasions on which an ad hoc compound is used to render a particular concept, but which compound can't be considered to be a satisfying translation.

I can illustrate the point with 'butikŝteli which both Wells and Benson give as a translation for 'shop lifting'.

Our knowledge of the world tells us that the analysis should be steal from a shop rather than the stealing of shops. So it is different from monŝteli, the stealing of money, and infanŝteli, (kidnapping or baby trafficking?).

However, from this compound 'butikŝteli, we can deduce nothing about the circumstances of the stealing . One is therefore at a loss to know whether 'butikŝteli' is shop lifting, or looting (as in the recent riots in England) or indeed the occupation of a robber who holds up corner stores.

Of course, usage could give the term a specific application, and quite possibly in German a similar compound word has acquired that specific meaning from usage. But as any Esperantist wil readily admit there are areas of life in which Esperanto is rarely used.

Of course, whether this vagueness matters or not, depends on your view of the function of Esperanto. If it is just a conlang hobby, it is hardly of any consequence. Nor if you see Esperanto as primarily a communication tool for the educated tourist.

But, if you think that Esperanto is suited for the purposes that any natural language is put to (a claim often made for Esperanto), it is not a trivial matter.

I think you should have posted this in the Esperanto forum rather than here. I mean, it's kind of an advanced topic, so surely it's more productive to discuss it in Esperanto with people from various countries, and not just rely on those that happen to speak English.

sudanglo:Whilst it is true that the meaning of many compounds word is fairly transparent from the components and a knowledge of the world, there are cases where the combinations are unsuitably vague in their application.

I think it is fair to say that combinations are often not completely clear by themselves and their meaning needs to be established by usage and dictionaries. As far as I known this is directly analogous in every language with similar compound words.

But what's the point? Are you doubting about the quality of dictionaries or about the maturity of the language itself? Or both. If a word like 'butikŝteli' appears in a dictionary with a specific meaning without real proof from reality, than we could say that that is a bad thing for a dictionary from a natural language. I'm not so sure about Esperanto though. It is after all a conlang, so a bit of prescription rather than description is perhaps less evil.

This phenomenon is good for illustrating true and false propaganda about the language; we can say that Esperanto is relatively easy because the components of compound words are a ready-made and natural mnemonic, but we can't say that complex words can always be instantly and easily made up as required, or deciphered for that matter.

sudanglo:It seems to me that there is a problem in principle with Esperanto's combinatorial structure

Among other things.

sudanglo:Whilst it is true that the meaning of many compound words is fairly transparent from the components and a knowledge of the world, there are cases where the combinations are unsuitably vague in their application.

In most cases, context eliminates ambiguity.

Why not just use "rabi" for your example and be done with it?

It's best to make it clear from the beginning that the meanings of compounds are suggested but not determined by the meanings of their elements.

sudanglo:I can illustrate the point with 'butikŝteli which both Wells and Benson give as a translation for 'shop lifting'.

Our knowledge of the world tells us that the analysis should be steal from a shop rather than the stealing of shops. So it is different from monŝteli, the stealing of money, and infanŝteli, (kidnapping or baby trafficking?).

However, from this compound 'butikŝteli, we can deduce nothing about the circumstances of the stealing . One is therefore at a loss to know whether 'butikŝteli' is shop lifting, or looting (as in the recent riots in England) or indeed the occupation of a robber who holds up corner stores.

In this instance, the more systematic pattern could easily be recovered by a slight adjustment of the target substantive: "butikumŝteli" would be "shopping-stealing", i.e. "stealing-while-shopping"); or "butikumindŝteli" would be "shopping-worthy-stealing", i.e. "theft of things worth shopping for".

I hesitate to suggest "butikumlevumi"...

I'm with TOdd on this one.

Some combinations have obvious meanings, and those cause no problems. I haven't done a survey, and so don't have any hard numbers, but my feeling is that this is true for most two element combinations of the sort that actually get used; I'm not talking about the probabilistic sea of purely random couplings - e.g., nucleus-cantaloupe - where even one meaning is questionable.

Where there are multiple possible interpretations of combinations, though, I think we have no choice but to simply communally "decide" in some fashion or another what the combination means. That act of deciding probably boils down to common usage. For those words where the term isn't likely to attract much usage in the first place, then whoever uses it likely will need to explain the desired meaning (perhaps explicitly, or maybe just through context).

I don't really see that as an issue, though, as it seems likely no language escapes this predicament. Otherwise, individual compound words would have to contain so much information that they would quickly become ridiculously (and therefore unusably) long, or they would have to be split into entire explanatory phrases - which, actually, we can do now, if we want.

sudanglo:I can illustrate the point with 'butikŝteli which both Wells and Benson give as a translation for 'shop lifting'.

NPIV2002 doesn't contain this word. I wonder if the 2005 version does? Maybe it just wasn't seen as being very important.

Chainy:

sudanglo:I can illustrate the point with 'butikŝteli which both Wells and Benson give as a translation for 'shop lifting'.

NPIV2002 doesn't contain this word. I wonder if the 2005 version does? Maybe it just wasn't seen as being very important.

It's not in PIV 2005, from a quick look at ŝteli. At first glance the word could refer to any of looting, shoplifting or a stick-up, so I'm not wild about it. But since Wells has it, I'll go along with it.