Why Esperanto is not my favourite Artificial Language
A comprehensive demolition of Esperanto’s pretensions, last updated 24 June 2002 by its creator Geoffrey A. Eddy. His last known data is Geoff’s homepage -> Artificial Languages -> Esperanto criticism A special rebuttal can be found at remush.be/rebuttal.
Obviously, this page is critical of Esperanto; Esperantists who aren’t comfortable with criticism of their language will thus read on at their peril. If you’re an Esperantist who’s stumbled across this by mistake and who is looking for some less objective material about the language, you’d be better off looking at the pro-Esperanto pages mentioned in the References section. Some devout Esperantists have been known to be upset by this critique; if you’re among them, you really ought to ask yourself why, and if it really matters.
Contacting the author
I take this opportunity to express my gratitude to all my correspondents, including some very polite and friendly Esperantists, who contacted me in reply to this page with corrections, criticisms and general comments. I’ll gratefully accept this kind of critcism, as long as it’s level-headed and objective; but I will IGNORE flamage and attempts by annoyed Esperantists to persuade me that IALs in general or Esperanto in particular are a Good Thing. Perpetrators of particularly lame flames will be named and shamed; if you can’t take criticism, you aren’t allowed to criticise in return.
If you want to convince me that I’m wrong in general, you’ll have to refute at least half of my points. Claims consisting of just one or two trivial “corrections”, as at least one correspondent seems to have believed was sufficient, will be ignored. Emails in Esperanto will also be ignored; I simply can’t speak the language well enough to unravel its complexities.
Oh, you want my email address. For anti-spam reasons it isn’t here, but you can email me from my homepage instead.
- Contents; this list.
- Introduction; why I wrote all this, and some myths of Esperanto debunked.
- Origin and development; how it came about and where it is now.
- Phonology and Orthography; Esperanto’s unappealing visual and aural aspects.
- Grammar; it pretends to be simple, but is it really?
- Vocabulary; a dismal failure to live up to the language’s promises.
- Syntax; a regrettable oversight.
- References; further reading and alternative viewpoints.
- Appendices; additional “material” which may be of of interest. Now on their own page.
This page is both a critique of Esperanto and a rebuttal of pro-Esperanto propaganda; it is not an attack on those who use the language, nor an advocacy of any other language as a world auxiliary. There are links to four other critiques in the References, so yet another one may not be necessary; however, during my brief period of interest in auxiliary languages I encountered enough annoying propaganda that I felt the need to write this page in response.
To start with, here are rebuttals to some of the myths which Esperanto propagandists like to present as fact. The remainder of this document, using some intelligence and a substantial amount of common sense, goes into considerably more detail.
Myth: Esperanto is naturally euphonious and beautiful, a masterpiece of constructed language design.
Reality: This is a matter of aesthetics and personal taste, and thus meaningless; for what it’s worth, my personal opinion is that Esperanto is straitlaced, rather uninteresting, and actually quite ugly, especially to look at. The very fact that it was intended as an auxiliary language, rather than a work of art, militates strongly against it having more than a very few of the features which make the best constructed languages so fascinating. And as for Esperanto being a conlang masterpiece, there are hundreds more interesting conlangs out there.
Myth: Esperanto is very easy to learn, hear, speak and use.
Reality: Unless you are familiar with at least two or three European languages, Esperanto will clearly contain many unnecessarily complicated and awkward features. The more European languages you speak, the easier you will find Esperanto; but the less you will then actually need it! In any case, the number of people who have learned Esperanto says nothing about how easy the language is to learn, since the vast majority of these have learned the language through choice and conscious effort. Rather, you need to consult those – especially monoglots from non-European linguistic backgrounds – who have tried but given up; a lot could be learned from their reasons why. The remarks in [RHD], especially those concerning the “eternaj komencantoj”, are relevant here.
Myth: The grammar and vocabulary of Esperanto are made up of elements common to many widely-spoken European languages.
Reality: A barefaced lie; many grammatical rules, and especially items of vocabulary, are derived from idiosyncracies common to at most a very few languages. For example, the plural ending – one of the most important grammatical markers – takes its form not from something internationally recognisable (such as the -s of widespread languages like English, French, Spanish and Portuguese), but the -j from one noun declension of Classical Greek.
Myth: Esperanto can reproduce the idioms of any other language exactly and without ambiguity.
Reality: No constructed language could ever do this, unless it was very complicated. Esperanto, in fact, ignores many subtleties of expression found in the natural languages it purports to be able to replace.
Myth: Esperanto is a consistently logical language, with a clearly-defined grammar free from nonobvious idioms, and thus allows you to express yourself clearly and unambiguously.
Reality: As you will discover later on in this webpage, Esperanto grammar contains many grammatical usages which are nonobvious, unstated, inconsistent or illogical. These provide plenty of opportunities for ambiguity or unnecesarily fine subtleties of meaning, and there are aspects of the grammar which Esperantists disagree over.
Myth: Esperanto is internationally neutral, and is the best (some say the only) candidate for an eventual world-wide auxiliary language.
Reality: Esperanto is overwhelmingly European in design, content and aspiration, and has nothing in common with most non-European languages – such as Arabic and Swahili, two other languages which are widely used as linguae francae in other parts of the world. It is thus woefully deficient for world-wide auxiliary use – a role which English, for all its faults, has been fulfilling perfectly well for decades.
Origin and development
Some history, which provides some explanation for what followed.
How it happened
[DH] has a good account of the history of the more successful auxlangs. For now, I’ll mention that Esperanto first appeared in 1887 as “Lingvo Internacia de la Doktoro Esperanto”, created by Dr. Ludwig Lazarus Zamenhof, an oculist who lived in what at the time was Tsarist Russian Poland. The pseudonym Esperanto caught on and has been used as the name of the language ever since. [I’ll be fair here: for all its failings, Esperanto does at least have a better name than many of its competitors. But a good name does not by itself a good language make.]
Esperanto was developed in response to Zamenhof’s unpleasant experiences of life, where there was much hatred between Russians, Germans, Poles and Jews. He believed that the hatred existed because the various peoples didn’t speak the same languages and couldn’t understand each other; he created Esperanto as part of a wider plan to overcome the linguistic barriers and improve cultural understanding, thus reducing or eliminate the potential for conflict.
While his aims were, and still are, laudable, his diagnosis was wrong; language is at most a symptom of cultural antagonism, rather than the actual cause, which is more likely to be a complex mixture of social, racial, religious and historical factors. It is highly doubtful that Esperanto would have prevented the American or Balkan civil wars, or the conflict in Northern Ireland, for example.
What is it for, anyway?
Some sources, including [TYE 1], claim that Esperanto is “in no way opposed to the national languages” and is intended for use alongside them. But others think otherwise and insist that Esperanto should be a universal second language, with national languages used solely within the countries in which they are spoken; once you’ve learnt Esperanto, you should never need (or even want!) to learn another language ever again. One correspondent even expressed the opinion that, because we have Esperanto, we don’t need specialist language schools any more.
Occasionally, one encounters claims that, because Esperanto was originally motivated by peaceful intentions, using it will automatically make one dedicated to bringing about world peace. Similar claims are made all the time about religions, and there’s an amusing counterexample: [WIRED] mentions that the US army used to use Esperanto as the language of enemy forces in mock battles.
Language or minority religion?
Here’s a somewhat subjective, but nonetheless provocative, viewpoint.
A correspondent with inside information made the very interesting remark that part of the attraction of Esperanto is the feeling of belonging to a minority religion – or, if you prefer, an organisation which thinks it has the answer to everything. Some Esperantists indeed promote the language with an aggressive quasi-religious zeal, implying that some sort of linguistic revolution is just about to happen and you are morally obliged to take part. Such an attitude is, of course, more likely to put people off than anything else.
On this subject, a correspondent who wishes to remain nameless tells me that “Esperanto’s popularity in non-European areas almost always has to do with various religions and cults that advocate some kind of globalism. For example, Spiritualists in Brazil, Baha’i in Iran and elsewhere, and Oomoto in Japan. There’s also groups like the Sennacieca Asocio Tutmonda (SAT, “World Non-Nationalists Federation”), who see Esperanto as a tool to join the world’s proletariat and incite socialist revolution.”. [Note the rather silly word for “non-nationalist”!]
The quasi-religious mindset is well exemplified by [PR], a text which has been described as “reeking of L. Ron Zamenhof”.
How many speakers?
The back cover of [TYE] claims “eight million Esperantists”, which is surely an exaggeration; the real figure is impossible to determine with any accuracy, but is generally believed to be between 50 thousand and one million. Most of these are in Eastern Europe, where the language’s origins ultimately lie, and Japan and China, where its principal attraction seems to be that it is not English.
Indeed, aside from on the Internet, I personally have yet to encounter a single speaker of it; and it’s a safe bet that in no country do as many as one percent of the population actually use it on a regular basis. Scotland, for example, has merely a handful of Esperanto speakers, the majority of whom are over fifty years old. The language has thus yet to succeed in its aim to be accepted as a global auxiliary, and I doubt it ever will; the vast majority of the world’s population regards it with overwhelming indifference.
Why Esperanto needs reforming
Zamenhof wisely never claimed that Esperanto was perfect, and he was initially keen to receive criticism and suggestions for improvements. Unfortunately, few people bothered to begin with, and he later declared that no changes could be made to the language; many Esperantists consequently like to claim that this proves that Esperanto was (and still is) perfect and has never needed changes of any sort at all. Ever since then, the desire to reform Esperanto has often been regarded as some sort of heresy, and any attempt to fix even the more obviously broken parts is doomed to failure.
Yet the language has been criticised ever since it first appeared. Many of its faults are clearly due to its method of creation and the prevailing attitudes of the time, in particular the late nineteenth-century “mechanistic” belief that all human language could be fitted into a fully logical and rational framework – “a sort of voiced Dewey Decimal System”, as Mark Rosenfelder succinctly puts it in his analysis of the twentieth century.
The grammars of Classical Latin or Ancient Greek have often mistakenly been upheld as models of logic and rationality, and they clearly influenced much of the design of Esperanto, such as the unnecessarily large number of word-classes (Arabic, by contrast, has three). This rather silly attitude is explicitly referenced in some early editions of Teach Yourself Latin; here’s a neat summary of it.
Moreover, because Zamenhof was a polyglot and not a linguist, Esperanto is not the elegantly designed lingua franca its supporters would have you believe; instead, as the vocabulary in particular bears out, it’s a composite of several European languages clumsily mixed together with some of Zamenhof’s own fetishes, but with little focus or guiding principles. In parts, such as the ridiculous spelling system, it’s radical where it should be conservative; in others, such as the overcomplicated grammar, it rather feebly compromises when it should be bold and radical.
Some of Zamenhof’s ideas were good ones to begin with, but one result of his lack of proper linguistic knowledge is that – frustratingly – all of them are compromised or botched in ways which would have been so easy to put right, without exception. It’s only fair to award him one or two points for trying, but he should also lose them for making so many silly mistakes in the process.
Undoubtedly, a properly usable international auxiliary language, designed with modern-day linguistic knowledge in mind, would be totally different from Esperanto – assuming one is actually possible, of course.
Phonology (sound-system) and orthography (spelling)
Despite all the claims of “natural euphony” or words to that effect, what Zamenhof did with Esperanto displays complete ignorance of any considerations of euphony, by any standards except perhaps his own. A book I found [in a second-hand shop] called “Step By Step in Esperanto” remarks that Esperanto sounds terrible if pronounced badly; surely a naturally euphonious language, if such a thing can exist, would not have this problem?
None of this stops [TYE 3] from presenting a highly questionable argument that, because Esperanto is supposed to sound like Italian, it must be “one of the most beautiful languages on Earth”. Any such resemblance, if not imaginary, depends almost entirely on the preference of both languages for words ending in vowels – a preference shared with Finnish, Swahili, Maori, Japanese and Chinese, to name but five.
The phoneme inventory
Common sense dictates that a properly euphonious world IAL should be easy to speak and to understand when heard; it should therefore have a small inventory of phonemes (contrastive sound-units) and definite rules restricting the combinations in which they may occur. The extreme is displayed by Polynesian languages like Hawaiian, Rotokas and Maori; all of these have about a dozen phonemes each and avoid consonant clusters altogether.
Zamenhof, however, is guilty of a lazy (and, in what is supposed to be an international langugage, unforgivable) error typical of beginning conlangers (see, for example, see my own first conlang): Esperanto’s phoneme inventory, as [JBR] shows, merely consists of the 34 phonemes which are apparent from the orthography of his native Polish dialect. This overlarge sound-system compromises the otherwise sensible decision to represent each phoneme consistently by its own letter; Zamenhof had to create the six extra accented letters shown below. From left to right, they are pronounced like the final consonants in “itch”, “edge”, “loch”, “rouge”, “ash” and the W in “wet”.
Laying aside my impression that these extra letters are just plain ugly, you have to wonder. Surely Zamenhof would have wanted his language to be as easy to disseminate as possible; but how many printing-presses and keyboards actually feature these letters normally, or did so in 1887? (Come to that, when did you last see an Esperanto typewriter?) And from a beginner’s point of view, these “new letters which look like the old ones” can only be confusing, slowing people down while reading and writing. Amusingly, some Esperanto journals still manage to forget the accents, or put them in the worng places [RHD]. And it’s interesting that part of the artwork on Radiohead’s album OK Computer misses the accent off the final J on a sign saying DANGHERA NAJBAR-EJO (“dangerous neighbourhood”)!
The extra letters have been given their own Unicode assignments in the Latin Extended-A page; however, not everyone has Unicode fonts installed (if you can’t see the extra characters here: ĉ ĝ ĥ ĵ ŝ ŭ, you probably haven’t). With this in mind, the best I can do to represent these extra letters here in HTML, without spattering bitmaps everywhere as some people recommend, is to use Zamenhof’s concession that the accents can be represented by a following H, with W in place of the accented U. Some people prefer to indicate all accents with following X’s, which is very nearly the least attractive way around a problem which should never have been there in the first place; to prove my point I need only borrow JBR’s example, cxirkauxajxojn (in my reformed spelling, tcirkawajoyn). This is a serious handicap; however difficult English may be to spell, it can at least be typed on any unaltered Roman-alphabet keyboard.
The vowels of Esperanto – the /i e a o u/ common to many languages – make up the only part of the sound-system which comes close to being sensibly designed; unfortunately, even this good point is compromised by an over-reliance on vowel groups and diphthongs. My confidence in [TYE] is not increased by its claim on the first page that “most national languages have twenty or more” vowel sounds – in fact, the global average is somewhere between 5 and 7. Even by the criteria of Esperanto, English has at most fourteen, and some languages (Arabic and Quechua come to mind) have as few as three.
As is typical of much of Esperanto, the consonant system is clearly a compromise between various European languages rather than something sensible, and the spelling system is similarly a clumsy mixture of several European orthographic traditions. Here’s the consonant system laid out in the traditional manner, using the Esperanto spellings:
Labial Dental Palatal Velar Glottal Stops p b t d k g Affricates c (1) ch gh Fricatives f v s z sh jh hh (2) h Nasals m n Liquids l r Glides w j
A little thought shows that neither this consonant system nor its orthography are satisfactory; the table has some surprising gaps (i.e. irregularities), a clear sign that Zamenhof didn’t really have much idea of what he was doing. (1) is DZ, a sound which most Slavic orthographies (including Polish) treat as a compound consonant (compare the Esperanto edzi “to marry”), while (2) is the sound of G in Dutch “negen”. And if the spelling system was really consistent and logical, JH would be ZH, as in most Roman orthographies for Slavic languages; CH would be KH, otherwise GH looks wrong (maybe JH?); and HH would be something else – or better, not there at all.
Indeed, the presence of both H and HH is one of the least defensible features of the consonant system; almost all of the Slavic languages treat them as the same consonant, and only Spanish of the Romance languages has either. There’s no reason why an IAL should need these sounds at all; [TYE 6] even has the cheek to describe HH as very rare and not worth troubling about.
One sound, two spellings
Esperanto’s orthography is supposed to follow the sensible “phonemic” principle of “one letter, one sound”, but what this amounts to in practice is that the internationally recognisable Latinate letters X Q are replaced by the Slavic compound spellings KS/KZ/GZ and KV. Weirdly, the principle is ignored by the single letters C CH GH, which represent compound sounds better represented by TS TSH DJH; Zamenhof’s Slavic bias is the obvious explanation. Thus, and contrary to [TYE 6] (“as a word is spelt, so it is pronounced, and vice versa”), some words with what to most people will sound like the same pronunciations can be spelt in two different ways; for example:
archata "bowed" artshata "appreciative of art" placpaco "place of peace" platspaco "flat space" sorchtrumpeto "magic trumpet" sortshtrumpeto "sock of destiny"
A possible reform
My first reaction to all this was, more or less, “yeuch! how horrible!”. It didn’t take me long to improve the spelling-system to get rid of all the accented letters [independently of [JBR]’s very similar approach, incidentally); this reform proceeds as follows:
- There really isn’t much point in an accent which is used on only one letter; why not spell the accented U (which comes from Belorussian) as W?
- There’s no need to write the affricates as single letters. Replace C by TS; you can now spell SH as C and CH as TC.
- There’s no harm in replacing J with the otherwise unused Y. This allows JH to lose its accent, and GH can now be more sanely spelt DJ.
- If HH is really necessary, the unused X (from the Cyrillic alphabet) can be used for it.
This orthography has the following advantages over Zamenhof’s:
- It’s slightly less ugly!
- Even though very few languages use C for SH, it’s no more confusing!
- There are no new letters to worry about!
- All compound sounds are represented by compound letters!
Regardless of how it’s spelt, however, Esperanto’s sound-system still leaves a great deal to be desired; it’s no coincidence that most of the awkward sounds in Esperanto are those represented by the accented letters, which represent distinctions of sound unknown to many non-Slavic languages. And have a look at the Esperanto section of this page with letter frequencies in various languages: the letters representing the most contentious sounds (with the exception of C) are all at the bottom, and HH W don’t appear at all. In other words, it really wouldn’t hurt to get rid of them altogether.
How the extra letters mess up the vocabulary
I’ve read more than once that the extra (accented) consonants are a Good Idea because they make it easier to compromise between different forms of certain words in different languages; thus, because “garden” is jardin in French and giardino in Italian, the Esperanto is ghardeno. But why not just make it gardeno (or the German equivalent, garteno) and do away with the GH altogether? Similarly, “stone” is shtono, a strange blend of German and English, whereas stono or steno would get rid of the accent.
Another supposed advantage of the accented consonants is that they allow some roots to be changed to avoid homonymic clashes with others; thus “post/mail” is poshto to differ from post “after”. This advantage is negated by Esperanto’s ability to create homonyms in other ways (see Appendix 2); and in this case, there are better words for “after” if you follow the principles of vocabulary-building like you’re supposed to – what’s wrong with malantaw, the opposite of antaw “before”?
Sounds in combination
Unlike proper languages, Esperanto has no explicit rules specifying which sounds may actually appear in combinations; appendix 1 uncovers what seem to be implicit rules. Nor are there any rules which simplify awkward groups of sounds to ease pronunciation; such rules aren’t allowed because of the necessity to retain complete regularity of spelling. Esperanto consequently suffers from problems of pronunciation which no naturally euphonious language would permit:
- Consonant clusters often approach unreasonable lengths: dekstra “right”, transskribi “to transcribe”, eksscio “former knowledge” and ekscio “flash of insight”, postscio “hindsight” (with medial STSTS!), altkreska “tall”, pendshnuro “hanging-rope”. [TYE 136]’s failure to define “easy to pronounce” doesn’t excuse these; not for nothing does “extra” appear in Japanese as ekisotura. You can even make up nearly unpronounceable words like lingvgvido, “language-guide”; this site has plenty more such, like malnovajhscienco, traktdistingo, ekzercmuziko and vibrnombro.
- Several initial clusters are also problematic: scii “to know” (which [TYE 45] admits is “difficult to pronounce”), knabo “boy”, skvamo “scale”.
- Groups such as those in ekzisti “to exist”, sveni “to faint”, akvo “water” are awkward for people used to assimilating for voice, such as (oddly) Slavs. kv is obviously different from gv (e.g. in gvidi “to guide”), and kz isn’t the same as ks in aksiomo, but why should we have to worry?
- It’s virtually impossible to pronounce the N’s in banko “bank”, panjo “Mummy” and senbarba “beardless” in the same way, especially if you’re talking quickly; so much for “one letter, one sound”.
- Tricky vowel clusters are not excluded, either: unuaeco “priority”, soifo “thirst”; while premii “to reward” differs from premi “to press”, and mato “mat” isn’t the same as maato “mate (on board ship)”.
Hungarian, Finnish, Turkish and Arabic are among the many languages which don’t allow any more than one consonant at the start of a syllable, while Hawaiian and Swahili – to name but two – don’t permit any consonants at the end of a syllable. Even Italian, which is more lenient, is still very careful about which consonant clusters it allows.
Esperanto, by contrast, allows up to three consonants at the beginnings of syllables and up to two at the ends, with no sign of any restrictions; think of the problems which monoglot speakers of one of the above languages would have with the five-consonant clusters in words like randstreki or transskribi. So much for not placing any particular linguistic group at a disadvantage.
About 75 uninflected words end haphazardly in consonants (detailed at the end of appendix 1); this is a small enough number to suggest that the system could very easily have been made totally consistent.
The strange case of J, W, and the diphthongs which aren’t
Diphthongs are groups of two or more vowels pronounced in one syllable, as in British English “I now toy”. Authorities disagree about whether Esperanto has them; my copy of Millidge’s dictionary (in the definition of diftongo, no less), says no, whereas [TYE 8] says yes. This is because Zamenhof stated (quoted in [DYER], “General Esperanto phonetics”) that the letters J and W are always consonants, whether before or after vowels, whereas vowels in groups are always pronounced separately. J and W thus contrast with I and U next to vowels, which is an unnecessary and annoying complication; consider the following pairs of words, in which the first has two separate vowels and the second contains what is, to all intents and purposes, a diphthong:
- naiva “naive”, najlo “nail”
- reiri “to return”, mejlo “mile”
- soifo “thirst”, sojlo “threshold”,
- kuiri “to cook”, tiuj “those”,
- praulo “ancestor”, frawlo “unmarried young man”
- reuzi “to reuse”, fewdi “to feud”
- iu “someone”, ju (a particle)
David Peterson pointed out that the position of the stress can distinguish these, but adding suffixes moves the stress and makes the distinctions much more difficult. It is characteristically sloppy of Zamenhof’s design that such contrasts exist, but are rare enough that they could have been got rid of altogether with little effort.
Note the comment on the same page of [DYER] that: “It was noted that out of 22 successive numbers of the British Esperantisto, six contained articles on the pronunciation of the aj, oj, uj. This would not have happened if the pronunciation were as easy for the English as claimed.” Few people are likely to perceive any difference between pairs of words such as rejri and reiri, or to be able to pronounce frawlo, mejlo and so on according to Zamenhof’s instructions without using diphthongs. Even worse, try pronouncing io and ijo differently!
The standard propaganda line, reproduced faithfully in [TYE 1], is that “… instead of the usual maze of rules [of grammar and syntax] … we have only sixteen short rules, which may be written comfortably on one sheet of notepaper”. Nobody has ever explained how a language with that little structure could be anything remotely resembling usable; indeed, the other 185 pages of TYE by their very existence clearly demonstrate that there’s a great deal more.
[Update: The proof of the pudding again! I’m informed that the plena analiza gramatiko de esperanto is 600 pages long – that’s about the same size as Thurneysen’s Grammar of Old Irish, a classic work on a notoriously complicated language. Apparently it even has an apology, in its foreword, for its length; even my Latin grammar is only 450 pages long. So much for “one sheet of notepaper”!]
The sixteen rules…
The 16 rules are shown below. Typically, they only tell you how to say things without explaining what it is you’re actually saying, nor what all the grammatical terms mean, nor indeed why you need to say things this way at all. As for the supposed “genius” of their alleged simplicity, any competent conlang creator could easily improve them, and indeed many already have.
- The only article is the definite, la, which is invariable.
- Nouns end in -o, plural -oj, in the nominative case. The accusative case is formed by adding -n to the nominative. Other cases are expressed by prepositions.
- Adjectives end in -a, and agree with the noun in case and number. The comparitive is formed with pli (adjective) ol, the superlative with plej (adjective).
- The numbers from one to ten are unu, du, tri, kvar, kvin, ses, sep, ok, naw, dek, and are invariable. Higher numbers are formed along the pattern of dudek unu for 21.
- The personal pronouns are mi, vi, li/shi/ghi; ni, vi, ili; oni; si for “I, you, he/she/it; we, you, they; one; -self”. The possessives are formed by adding -a.
- The indicative verbal endings are -as -is -os for present, past and future tenses. There are corresponding active participle endings -ant- -int- -ont-, passive participle endings -at- -it- -ot-, the subjunctive ending -us, imperative ending -u and infinitive ending -i.
- Adverbs end in -e and compare in the same way as adjectives.
- All prepositions govern the nominative case.
- All words are pronounced exactly as spelt; there are no silent letters.
- The stress accent is always on the penultimate syllable.
- Compound words are formed by simply joining the root words; the chief word stands at the end.
- If a negative word is present in a clause, ne “not” is left out.
- Motion towards is indicated by the accusative case.
- Every preposition has a clear and precise meaning. Je is an indefinite preposition which may be used when no other preposition would express the meaning adequately. Instead of je the accusative case may be used.
- Foreign words do not alter their pronunciation, but are re-spelled according to Esperanto’s rules. It is preferable, however, to build up the word from Esperanto’s own resources.
- The final letter of nouns and the article may be elided for reasons of euphony.
… and what they don’t tell you.
Much of the grammar may seem to be a welcome simplification of the sophisticated grammars of Classical Latin and Ancient Greek, or even of most modern-day European languages, and ignorance of non-European languages other than these is doubtless why much propaganda claims that the grammar is the simplest achievable. However, considerable further simplification is possible, especially when it is borne in mind that many non-European languages indeed have a simpler structure than Esperanto (see [RHD] for a good quote). Here’s a short list (it could easily be twice as long) of some things Zamenhof either didn’t know or tacitly ignored.
- Articles are actually pretty rare in the world’s languages; to name but a few, Finnish, Swahili, Japanese, Chinese, and most Slavic languages all do without. Millidge’s dictionary claims that “the use of the article is the same as in the other languages”, which is complete nonsense since the uses of articles differ from language to language. How many languages say the equivalent of la kvar “the four” when telling the time?
- What are “the other cases” referred to in rule 2, how are they used, and why are they important enough to deserve a mention? The usual answers (“the genitive is expressed with de“, etc.) betray what seems to have been a nineteenth-century assumption that classical grammar is a constant of nature, rather than a fluid and more or less accidental convention; grammatical case is no more necessary than grammatical gender. The kazo akuzativo is examined in detail later on; and surely there are better plural endings than the unsightly and awkward -j? (See the introduction for where this comes from).
- No less an authority than Zamenhof himself is on record as conceding that agreement between the adjective and noun is unnecessary (“superfluous ballast”, in his own words in 1894), and indeed there’s no good reason why you should have to say grandaj hundoj “big dogs”, la hundoj estas grandaj “the dogs are big”, and mi vidas la grandajn hundojn “I see the big dogs”.
- Note the oddity that the word for “one” has two syllables while the rest have just one; and which part of speech do numbers belong to, exactly?
- Why is the pronoun system nothing more than a copy of the English one, when something else would surely have been more useful? Why, for example, is there no pronoun meaning “he or she”, and why is gender only differentiated in the third person singular? (David Peterson refined my original point here.)
- The verbal system may look straightforward, but the grammar doesn’t mention that you can form no less than 36 compound tenses with the various tenses of esti “to be” and the participles. This is far too many, and many of them are likely to confuse speakers of (for example) many Asian languages, which manage well enough with something much simpler. And if subjunctives, future tenses and participles are really necessary, why are there no “subjunctive participles” like vidunta? And is it a subjunctive mood, a conditional tense, or something else?
- Dutch and German get along fine without worrying about the distinction between adjectives and adverbs: what else could “love me slow and tender” possibly mean?
- Most prepositions can actually govern the accusative case, too; see rule 13. Esperanto is burdened with far too many prepositions, many of which also do the work of conjunctions and adverbs, and I’ve never seen a definitive list of them all.
- I’ve dealt with this rule in depth above; it has nothing to do with grammar or syntax!
- This rule, which is curiously the same as in Polish, has nothing to do with grammar or syntax either. The rigidity of the stress causes some distortions: why should words like nacio “nation” stress on the I, instead of on the first A as in every other language which contains the word? And do long words have any other secondary stresses?
- Another rule of grammar which has nothing to do with grammar! The problems of compounding are dealt with in the Vocabulary section.
- This rule is pointless and makes no practical difference to the language: multiple negatives are common in many languages, for example “I don’t know nothing” in colloquial English.
- This rule would be unnecessary if the uses of prepositions had been better thought through; see the section on the accusative case.
- Counterexample: [TYE 176] provides six meanings for de, and helpfully points out that la amo de Dio “the love of God” is ambiguous: is it “God’s love” or “some entity’s love of God”? Note too malamikoj de la urbo: is this “enemies of the city” or “enemies from the city”?
- Yet another rule which has nothing to do with grammar or syntax, and it doesn’t explain what to do with foreign words which have sounds not resembling any used in Esperanto. Some of Zamenhof’s less fathomable creations, e.g. ghis “towards” from French jusque, suggest that you either guess or change them at random – which rather defeats the whole point of Esperanto being an “internationally recognisable” language.
- Rule 16 is too trivial to deserve comment.
What this means is that eleven of these rules are unnecessary, and numbers 2 through 6 are open to further simplification: for example, do you really need all those confusingly similar verbal tense forms, when temporal adverbs would do just as well? (Cf Welsh rydw i’n wedi siarad “I am after talking”.)
More general comments
Esperanto propaganda and teaching guides place great importance on the principle of marking a word’s part of speech by its ending, even though it isn’t followed consistently; for example, numbers and prepositions have no consistent endings, while pronouns take the same terminations as verb infinitives, and the correlatives have a system all of their own.
The principle, despite its appearances of logic and rigour, is in any case of debatable merit, since numerous otherwise internationally-recognisable words have to be mutilated to make them fit into the language, with unpleasant results. Proper names come off worst; Edimburgo is just about acceptable, but Mario (yuck) for “Mary” is really not on; and nobody would realistically want to refer to the star Alpha Lyrae as Vego, when it’s always been called Vega! Even a name like Asia – recognised the world over and stressed on its first /a/ – has to be mutilated to Azio and stressed on the /i/.
Moreover, Zamenhof was unaware of one of the principal features of all human language: redundancy. Esperanto is sometimes lauded for its succinctness, whereby it can express things more economically than most other languages; however, this in practice equates to a greater density of information, with a correspondingly greater likelihood of misunderstanding as a result of mistakes – especially since much of the grammatical information is carried in the unstressed final syllables which mark the parts of speech. The best examples of this are the pronouns and the very similar verb tense endings, which in a properly designed language would be easier to distinguish.
In any case, the claims of succinctness work both ways: la hundo de la viro takes five words and seven syllables to say what Gaelic and English say in three of each: ců an fir, “the man’s dog”.
The accusative case
The accusative case is without doubt one of Esperanto’s least necessary features, and thus one of the most heavily criticised. It seems to exist principally as a concession to classical grammar (and thus to boost Zamenhof’s credibility with nineteenth-century academics?), but the language would be far better off without it. It’s supposed to free up word-order, which according to apologists is important for poetry and literature; but surely basic ease of communication, without having to worry about the finer points of grammar, matters more? Distinct accusative inflections have disappeared from many languages in the past; even in German, only masculine singular nouns have them; and neither the Chinese poets nor Shakespeare had any problems without them.
All would be well if, as in Ido, the accusative case was only obligatory when resolving ambiguities; surely Esperantists don’t get confused with sentences like la pomo mi manghas “I eat the apple”? Unfortunately, it’s also used in several nonobvious and idiomatic – i.e. unpredictable – ways; [TYE 66] claims that it can be used in place of any preposition, a rule which is asking for trouble even if you don’t take it to mean that it effectively renders all prepositions equal in meaning. Thus you get strange constructions like du futojn alta “two feet high” and jhawdon la 1an de decembro “Thursday 1 December”, all of which could be more clearly expressed differently. You can even add the suffix to adverbs, a rule which doesn’t appear in the grammar (thus alten “upwards”); yet many words, including the article, numbers and some correlatives, don’t or can’t take it at all.
According to one of my correspondents, who seemed very sure of the point, the accusative case can only replace je; somebody is fibbing somewhere! The confusion between the accusative case and je, which is officially blessed in rule 14, gives rise to a curious ambiguity. A commonly mentioned example of the use of je is veti je chevaloj “to bet on horses”, which can also be veti chevalojn. So, since veti monon is correct for “to bet money”, veti monon chevalojn is quite reasonably both “to bet money on horses” and “to bet horses on money”!
Rules 8 and 13 of the grammar mean that en la domon “into the house” and en la domo “in the house” differ in the meaning of the preposition, but express this difference by changing the noun. This distinction does not extend to any other types of motion; thus el la domo “out of the house”; and if compound prepositions like de sur “off” (i.e. “from on”) are permitted [TYE 50], what’s wrong with the entirely unambiguous al en la domo?
Zamenhof’s unnamed part of speech
The ending -aw appears on – to the best of my knowledge – a mere 21 words, suggesting that Zamenhof once tried to create a new part of speech of unidentified function but gave up without tidying up the mess. Many of the words are both conjunctions and prepositions (malgraw “despite”, antaw “before”, anstataw “instead of”), but some are one or the other (aw “or”, cirkaw “around”), a few are bona fide adverbs and should thus end in -e (baldaw “soon”, apenaw “hardly”), and a couple are neither (adiaw “goodbye”, naw “nine”).
Not mentioned in the grammar, but vital to the language anyway, are 45 “correlative” words which are formed by joining together one of 5 prefixes to one of 9 terminations; thus i-o “something”, neni-u “nobody”, chi-a “all kinds of”, ti-el “thus”. This is superficially one of Zamenhof’s best ideas, and it looks clever enough to have persuaded some Esperantists that it’s some sort of an indication of genius; but simple and transparently obvious phrases like de tiu “that one’s”, tia ejo “that place”, and so on would be far better than arbitrary words which have nothing to do with the rest of the language. In any case, Zamenhof entirely typically contrived to make a mess of it, turning a potential silk purse into yet another sow’s ear:
- There’s no prefix for “this”, so you have to put chi before “that” to get the horribly contrived chi ti-. Yet the important word “now”, which should thus be the awkward chi tiam, is actually the entirely arbitrary nun.
- The system of grammatical endings insisted upon in the grammar is, for no obvious reason, completely ignored; for example the adverbial ending is now -el (ki-el “how?”), while -e (ki-e “where?”) signifies place.
- Possession is indicated by -es; thus ki-es “whose” is distinct in form from de “of” (used with nouns) and -a (used with pronouns). That’s three different, non-interchangeable, ways of expressing the same grammatical relation – precisely the sort of difficulty Esperanto is supposed to have eliminated!
- The rather pointless distinction between the endings -u and -o doesn’t apply elsewhere.
- It would be nice to differentiate question words from relative pronouns, rather than lumping them together under k- and creating ambiguity in sentences like mi diris al la homo kiu parlis “I told the man, who spoke (?)”.
- Are the correlatives which end in consonants immune from inflection? Are you, for example, really supposed to say mi havas neniomn for “I have none”?
Some strange words result from inflecting certain correlatives, such as iujn, neniejn; and of course, you have to say kiuj estas ili, with plural correlative, for “who are they?”.
Bizarrely, you need to use the correlatives in comparisons of equality: mi estas tiel inteligenta kiel vi “I am as intelligent as you”, rather than something analogous to mi estas pli inteligenta ol vi “I am more intelligent than you”, which would surely be clearer and more obvious.
The introduction to Millidge’s dictionary contains a passage which implicitly berates English-speaking Esperantists for not being able to use the language properly, using the following two quotes from Zamenhof in Unua Libro de la Lingvo Esperanto to make the point:
“… everything you write in the international language will be clearly understood at once (with a key or without it) by anyone, who not only had never previously mastered the grammar of the language, but had never even heard of its existence” (italics mine).
“… anyone learning my language without a text-book (…) will not imagine that the structure of the language differs in any way from that of his own native tongue”.
Both of these amount to little more than wishful thinking, since neither remotely approaches the truth; you need look no farther than the grammar, which quite clearly differs in structure from just about every language on the planet, to refute the second. These quotes look particularly laughable in the light of Esperanto’s vocabulary, which despite Zamenhof’s efforts, and the insistence of his followers, constitutes by far the most baffling and confusing part of the language.
Esperanto’s vocabulary displays nineteenth-century mechanistic ideology in full flourish. The underlying assumption – inspired by Esperanto’s predecessor Volapük, as Zamenhof openly admitted – is that every word ever spoken in every language can be converted to an unambiguous and unique combination of “roots”, which express basic meanings, and “affixes”, which modify them; and by keeping the number of roots to a minimum, the memory-load is kept down, and a careful choice of affixes compensates by adding expressive power to the system. However, neither Zamenhof nor Volapük’s creator appreciated that meaning – like grammar – is in practice fluid and largely unpredictable, and unsuitable for shoehorning into such a rigid system. The English words “silly” and “villain”, for example, once respectively meant “happy” and “farm worker”.
According to [DYER] (penultimate paragraph of “Adjectival suffixes”), Zamenhof originally attempted to build up Esperanto’s entire vocabulary with less than a thousand roots; later on, because this number was patently inadequate, it was expanded to nearly three thousand, but even this turned out to be not enough. A rather telling consequence of this is that “… there is hardly an Esperanto translation of length which does not contain some words not to be found in the official vocabularies” [same source].
The result of all this is that Esperanto’s vocabulary reproduces much of the very same chaos of natural language which it is supposed to have eliminated. Despite claims like those of [TYE 2] (the “regularisation” of the vocabulary is “regular and complete”), the reader will soon discover that the meanings of many Esperanto words are idiomatic, illogical or inconsistent with others, or even with languages with which a speaker may be familiar.
In any case, there’s a fundamental problem with such a vocabulary-building system: any potential gain in the reduction of the memory-load is offset by the necessity of having to work out what the words are supposed to mean, even without considering all the exceptions, irregularities and idiosyncracies. A communication on the auxlang mailing list to a learner a while ago gave it away: “Don’t learn the roots, learn the words”. The proof of the pudding, as they say…
[TYE 2] claims rather inaccurately that the roots come from “the most important languages of Western civilisation”; in fact, working out the exact origin of any given Esperanto root reveals much about the essentially arbitrary nature of what is supposed to be a “logical” language. The sensible thing would have been to choose roots which would be recognisable to the greatest number of people; however, Zamenhof’s choices often seem to be as good as random. The only obvious guiding principle seems to have been a desire to take obscure roots from as many languages as possible, probably to avoid perceptions or charges of bias.
Many roots – those favoured by propagandists when giving examples, strangely enough – are imported more or less without major change, such as hundo “dog” and vivi “to live”. However, many more roots are unrecognisable for one reason or other, such as those which retain the same spellings as in the source languages and become unrecognisable when spoken according to Esperanto’s rules; for example:
word meaning source boato boat English (gains two syllables) birdo bird English (sounds more like "beard") soifo thirst French (sounds completely different) ohmo ohm German surname (the H has never been pronounced) komenci to begin French (gains a /t/)
or the roots which are distorted in nonobvious ways from their forms in the source language, among which are:
word meaning source boji to bark French "aboyer" lerta clever "alert"? fulmo lightning "fulminate"! yes, really! venko victory French "vaincre" mejlo mile English pordo door Romance aldo alto Italian
Investigation reveals that the last two were altered to avoid confusion with derivations from alta “high” and porti “to carry”; there are surely better words anyway, since “alto” is not really an everyday word for most people.
A few roots even undergo both of these Zamenhofian transformations and become confusing twice over. If you don’t know its Latin origins, one such is kvieta, which neither looks nor sounds like its English equivalent, “quiet”.
The random polyglot origins of the root-stock unavoidably create many “false friends”, i.e. words with the same form but different meanings in two or more languages. This means that many Esperanto words don’t mean what you might at first think; thus monoglot English speakers would have trouble recognising the actual meanings of flava, flata, logo and farti (respectively “yellow”, “flattering”, “enticement”, “fare”), not to mention the bizarre contrivance adulto “adultery”. Even something as basic as “to read” is less than obvious; a speaker of English unfamiliar with Romance languages who encounters lego (the derived noun) would have to consult several words (including “leg”, “legal”, “Lego”, “legible”, and “legless”) before being told that it came originally came from Latin.
Some meanings are expressed by more than one root: “master” could be any one of estro, mastro, or majstro; “colour” is hhromo in hhromotipo but natively farbo or koloro; and [RHD] mentions that all three of redakt-, redaktor- and redakci- are “commonly used” to express “the idea of editing”. And a few roots are completely unnecessary: keno “resinous wood” is perhaps the best example, especially since it has a much better equivalent rezina ligno, leaving keno for something much more useful. We may never know why Zamenhof felt the need for separate roots (kiso and shmaco) for two different sorts of “kiss”.
Making new words
Before you can derive words from any given root, you have to worry whether the root is inherently nominal (shton-o “stone”), verbal (vid-i “to see”) or adjectival (blank-a “white”); and there are plenty of inconsistencies here to trip you up: for example, martelo “hammer” is nominal, but shoveli “to shovel” is verbal. You also have to know if a verbal root is transitive or intransitive, i.e. whether it takes objects or not when not otherwise suffixed, since the distinction is unnecessarily important for the syntax.
Having chosen your root, and reassured yourself of its category, you can make new words from it by a combination of three methods, which I’ve chosen my own names for since there doesn’t seem to be any accepted terminology:
- Conversion – adding a grammatical part-of-speech ending directly.
- Compounding – joining two or more roots together.
- Derivation – adding one or more affixes before the grammatical endings.
As with the grammar, you’re only told that these methods make new words, but there are no rules to help you work out what the words are supposed to mean. This is a surprising omission in a language which claims to be “logical” and “precise”, and it jars with Zamenhof’s remarks at the beginning of this section.
A thorough overhaul of the vocabulary-building principles was carried out in Ido, the most successful of the many reforms of Esperanto; the principles are explained in [DYER], and it’s instructive to observe how Ido’s vocabulary is so much clearer, more consistent and regular than Esperanto’s. (This should not be taken to imply that I advocate Ido as la internaciona linguo, merely that I consider Ido to be a considerable improvement on Esperanto.) Even more thorough in combining precision and clarity with expressive power and logic is Rick Morneau’s admirable monograph on Lexical semantics, the rigour of which shows up the inadequacy of Zamenhof’s efforts.
A mere change of grammatical ending gives no indication of the number of derivational processes a word has undergone, nor of their nature; Ido’s sensible “principle of reversibility” fixes this rather nicely. The following non-obvious results of conversion show what can happen:
Original word Derived word blanka "white" blankas "is white" fundo "bottom" funde "thoroughly" shteli "to steal" shtele "furtively" suno "sun" suna "solar, sunny" (a sanctioned ambiguity) vivi "to live" viva "alive, lively" (another sanctioned ambiguity) hundo "dog" hunde "like a dog? doggy-style?" legi "to read" lega "???"
Consider shtono “stone”; the converted adjective shton-a could mean any of “stony”, “made of stone”, “made of stones”, “like stone”, or possibly something else again. And there’s no way to work out what the corresponding verb shton-as means, since you don’t know if it’s derived from the noun (“is a stone”?, with one step of conversion) or adjective (two steps, meaning something like “is stony”, “is made of stone”, “is like stone”). It could even be “stones”, as a punishment for adulto.
And consider the arbitrary and idiomatic adjectives sata “satiated” and flata “flattering”. These seem to be shorter synonyms for the more precise participles satita and flatanta, from the verbs sati “to satiate” and flati “to flatter”; yet the adjectives have the same form, whereas one participle is passive and the other active. This is not “regular and complete” by any criteria I recognise.
It’s no surprise, then, that an unnamed Esperanto textbook, quoted under “Derivation” in [DYER], recognises the problem and admits that “… it follows that the exact meanings of [words] have to be looked up in the dictionary…”.
The most glaring defect in the word-building system is the absence of any rules which explain what happens when you combine roots together. The dictionaries thus contain many compound words with meanings not predictable from their components; for example, postvivo is ambiguous between “survival” and “afterlife”, and fruktodona (also the almost unpronounceable fruktdona) “fertile” and poshomono “pocket-money” are formed idiomatically.
Strangely, the dictionaries also give many words which aren’t built up from Esperantine roots at all; many of these words are Latin or Greek compounds with elements which would be more recognisable than their Esperanto equivalents. “Astronomy” is thus astronomio – a form reasonably obvious to everybody – and not the rather ugly Esperanto compound stelscienco, while mikroskopo “microscope” would provide the useful root mikra- for “small” in place of the nasty malgranda. prognozo is used in [TYE] for “forecast” as in “weather forecast”, but some equivalent of “prediction” would be more accurate.
The individual elements of these international words sometimes clash with existing Esperanto roots, making a mockery of the entire system. The Fundamento, one of Esperanto’s sacred texts, has this problem: fund- has something to do with bottoms or thoroughness, and ment- is “mint”, so what should presumably mean “the Fundamental” actually means “mint of the bottom” or “thorough mint”!
You’re supposed to stick to the “official” affixes and not make up your own to remedy defects in the language; yet, as with the prepositions and conjunctions, nobody seems to have a definitive list anywhere. For example, [TYE 190-1] gives 10 prefixes and 31 suffixes, but other sources suggest different numbers.
It’s not obvious why derivation is always treated as being distinct from compounding, since affixes are elevated to the status of roots by the rule which states that an affix may be used as an independent word when used with the appropriate grammatical ending (for example, ar-o is “collection”). All the books I’ve read quite clearly make the distinction, however, and I feel obliged to defer to them. Yet the distinction is actually meaningless in practice, and Zamenhof’s hybrid system would thus be better replaced by a single set of properly recognisable roots.
Another problem with Esperanto’s affix system is that, like the choices of parts of speech and roots, it is based upon an essentially arbitrary set of criteria. It’s debatable whether it is possible to choose a universally useful set of affixes on purely objective grounds; Zamenhof’s affixes are idiosyncratic and all questionable in one way or another. Moreover, according to Lexical Semantics, a logical system of derivational suffixing is only really possible with verbs; most of Esperanto’s affixes, by contrast, are principally nominal.
Affixes the language doesn’t need
All of the following affixes are unnecessary and would be better expressed by separate words.
- mis- “wrongly”.
- fi- and -ach-, two derogatives; but there are no affixes with the opposite meanings. Why?
- eks- “former, ex-“.
- ek- “sudden or momentary”; this is easily confused with eks- when applied to roots which begin with s-. This has a place as a verbal aspect marker, but as such it would be better as a suffix.
- -ing- “holder for one object”, which has no relation to teni “to hold”.
- re- reproduces the ambiguity of its Latin source, which means both “again” and “back”, creating words with two different, sometimes almost opposite, meanings: re-skribi “to reply” or “to rewrite”; re-iri “to return” or “go again”.
Affixes with unnecessarily vague meanings
-ar- creates arb-ar-o for “forest” (“tree-collection”), which could also mean a line of Lombardy Poplars. Less forgivable is the misleading word ov-ar-o “collection of eggs”, which pointlessly duplicates the meaning of nesto “nest”.
-uj- “container” really shouldn’t be used to make names of countries such as Skot-ujo “Scotland”, nor is -ej- “place” justified in words like lern-ejo “school” and pregh-ejo “church”. These last two words are literally “learn-place” and “pray-place”, which are too general in meaning; they could equally well refer to many other things such as “classroom” and “prayer room” in a school building.
-ajh-, “something made from or possessing the quality of”, is possibly the vaguest; it gives rise to idiomatic oddities like akr-ajho “edge” from akra “sharp”, ov-ajho “omelette”, ter-ajho “soil” from tero “the Earth”, korp-ajho “flesh” from korpo “body”, and others in [TYE 77-8]. It also creates pairs of words which pretend to have different meanings but don’t; thus both kava and kav-ajha are given as “hollow” in my dictionary.
Affixes of questionable benefit
mal- doesn’t mean “badly” or “wrongly”, but forms opposites – a device Esperanto overuses to a ludicrous extent, for which reason it’s probably the most hated affix in the language (certainly by me!). Thus common words like “small, short, narrow, old, left, bad, different” have to be mal-granda, mal-longa, mal-largha, mal-juna, mal-dekstra, mal-bona, mal-sama; and “loud” is the ridiculous malkvieta. Not only do these words require unnecessary mental gymnastics, they also gets monotonous if you have to use more than one or two of them. Even a basic meaning like “to open” is not exempt; it’s mal-fermi, i.e. the opposite of fermi “to close”!!! As ever, there are unexplained exceptions: “left” and “right” are opposites (dekstra, mal-dekstra), but “north” and “south” aren’t (norda, suda); why? And David Peterson informs me that some people like to say trista for “sad” anyway, rather than the malfelicha you’re supposed to use.
The augmentative -eg- and its opposite -et- reduce many possible degrees of size to just three. Thus the triplet vento, vent-eto, vent-ego “wind, breeze, gale” replaces the entire Beaufort Scale, and arb-eto (from arbo “tree”) turns out to be “small tree, shrub”, requiring the desparate-looking contrivance arb-et-ajho for “bush”. Note also the typically idiomatic derivation rid-eti “to smile” from ridi “to laugh”, which is clearly a lame attempt to keep the number of roots down; it would better mean “to chuckle”.
eta, derived from the suffix, seems to be a synonym for malgranda “small” – but if it isn’t, as many sources imply, why is the distinction necessary? Is mal-eta the same as ega? Can you use -et-eg-a and -eg-et-a to make finer distinctions of size? Together with the vagaries of derivation and conversion, these suffixes provide further scope for ambiguities: if rugh-eta (derived from an adjective) is reasonably “reddish”, then shtoneta (derived from a noun) is equally reasonably both “a bit like a stone” (shton-eta) and “like a pebble” (shtonet-a).
And, for a language with supposedly high ideals and no grammatical genders, there’s no excuse for the excusively feminine suffix -in-, which requires “woman” to be vir-ino “a female man” (not, strangely, the more neutral homino “female human”); the hypothetical converse, fem-ula for “man”, is equally absurd.
Inconsistently used affixes
-il- is the worst offender: an Esperantist kombas “combs” with a komb-ilo and razas “shaves” with a raz-ilo, but brosas “brushes” with a broso, not the entirely logical bros-ilo (which is what, exactly?). It’s also too vague: paf-ilo “gun” could mean anything which shoots, from a peashooter to a cannon; indeed, as [RHG] mentions, the official word for “cannon” (paf-il-ego “big shooting tool”) is often rejected in favour of the more precise kanono.
-ec- “quality” is necessary to make abstract nouns from nominal roots. Thus homo, hom-eco “man, manliness”; but compare the inconsistent firma, firmo “firm, firmness”. blanko, blank-eco probably both mean “whiteness”; a correspondent informs me that blanko is used in phrases such as “the white of the eye”, for which something like blankajho would be better.
-an-, -ist- and -ul- all represent various types of people; note the inconsistency with mistiko “mysticism”, mistik-ulo “mystic”, but katolik-ismo “Catholicism”, katoliko “Catholic”. (There are further perils here: you might think that katoliko could be a compound with kato “cat”, before consulting your dictionary and discovering that liko doesn’t actually mean anything.)
Brendan Linnane points out that the suffix -on-, which is used to form fractions (e.g. ses-ono “a sixth”), is also used on the word for “million”, miliono, which is not a fraction; note its similarity in form to “thousandth”, which is mil-ono, and in sound miljono, which could be anything.
Because the affixes are short and arbitrary, many of them appear as parts of longer roots and so give rise to words with several possible meanings. An example for now is sukero, which means both “sugar” and suk-ero “a drop of juice”; more such words may be found in Appendix 2.
Further ambiguities also arise when you mix affixes together, since there is no indication of what affects what. The classic example is mal-san-ul-ej-o, ultimately from the root san- “health” with the affixes mal- “opposite”, -ul- “person” and -ej- “place”. You’re supposed to work out that this means “hospital”, literally “place for a person the opposite of well”; even with this derivation it could also mean “private hospital room”, “epidemic zone”, and so on. If instead you parse it as malsan-ulejo, you get something like “sick building syndrome”. Likewise, malgrandeta is both the opposite of grandeta “largish” and the diminutive of malgranda “small”.
There are at least nine ways of constructing something which looks equivalent to English “different”, but probably isn’t: alia, malsama, nesama, malsimila, nesimila, neidenta, malidenta, neegala, malegala.
Things which seem to be affixes but aren’t
The four words debato, debeto, debito, debuto may seem to be related derivations from a stem deb-, but they aren’t. Worse, some affixes mean different things at different times; thus the prefix eks- “former” has its meaning changed to “out of” (which should be el) in words such as eksciti “excite”, ekstrakti “extract”, ekstrema “extreme” and eksporto “export”.
Similarly, many words begin with pre-, which seems to mean “before”, however there is no such prefix; the actual Esperanto equivalent is antaw-, which should really have been left as ante-. And a lot of words derived from Latin begin with kon- or its assimilated form kom-, retaining its meaning of “with” for which the Esperanto is actually kun; the unwary reader or listener must therefore wonder if the word is a compound with some form of koni “to know”, or perhaps komo “comma”.
Affixes which aren’t there
The derivative apparatus is deficient in other ways too; one obvious omission is an affix meaning “the result of an action”. Thus the nearest to “a piece of writing” or “something written” seems to be somewhere between skribitajho or skribajho, but the usual meaning for -ajh- doesn’t imply this. Another try is skribito, but this properly means “a person who has been written”, which is nonsense even in Zamenhofese. There’s always skribo, but that could be something else again; although, in Ido, we can be sure that it’s what we’re looking for.
The absence of any rules governing word-building in Esperanto and of any guidelines for working out the meanings of the results mean that, in practice, claims about the “expressive power” of the vocabulary are really inducements to be clever at the expense of clarity. As the great Otto Jespersen remarked in [JESP] (I have reformatted the text slightly, but not change the meaning):
“When in an Esperanto book one stumbles on the word ghistiamajn and succeeds in making out that it means ‘previous’ and is a compound of the following elements: ghis […] ‘up to’, tiam then, -a adjective ending, -j plural, -n accusative, then one cannot help asking oneself in the face of so much ingenuity if it is really necessary for an auxiliary language to be made up of such utterly arbitrary elements – a question to which the whole of the subsequent history (and of the second part of this book) gives an emphatic answer in the negative.”
Why would anyone want to say that for “previous”, anyway, when the word-building system allows something like antawa? It isn’t pretty, but at least it’s reasonably transparent.
In Esperanto you can create false friends to your heart’s content, and in the process destroy any vestiges of “international recognisability”. Here are a few:
Word Meaning What it resembles but doesn't mean fingr-ingo thimble finger ring fos-ilo spade fossil for-esto absence forest ses-ono one sixth season virgul-ino female virgin female comma (from French "virgule")
virgulino has several other possible but incorrect meanings, such as a male gulino (by analogy with virbovo “bull”), or a hermaphrodite gulo. It’s actually formed from the adjective virga “unspoiled” – which is also used of, for example, unploughed fields.
[TYE 1] claims that syntax is just word-order; it’s actually the set of rules which say what words mean when used in combination according to the rules of the grammar. Definite rules of Esperanto syntax are surprisingly hard to find, which makes a general criticism difficult beyond observations that it typically seems to be very European and thus problematic for non-Europeans. Nowhere, for example, are you told how to form questions and relative clauses, use capital letters and other punctuation, and so on; you’re supposed to know that by intuition (or perhaps a classical education?). Instead I will provide a few specific results of Zamenhof’s failure to consider this important subject properly.
Example 1: consider the sentence estas shtelata la hundo de la viro, literally “is stolen the dog by/of/from the man”. Not only is the meaning of the preposition de ambiguous in several ways [TYE 176], but it’s not obvious whether the first two words mean “is being stolen” or “has been stolen and still is”. Thus this innocent-looking sentence can mean at least six completely different things.
Example 2: you can’t say both la domo brulas “the house burns” and mi brulas la domon “I burn the house”, since the verb is intransitive (i.e. taking no object) in the first sentence and transitive in the second. Instead, since the root brul- is intransitive, you have to make it transitive by adding the suffix -ig-, regardless of the fact that the very presence of an object – marked, moreover, with the mandatory accusative suffix – is doing just the same: mi bruligas la domon. Conversely, to satisfy the Esperantists that by not providing an object you really mean to use a transitive verb intransitively, you have to use the suffix -igh-, and don’t mix up your accents! These two suffixes have their uses with adjectival suffixes, viz. blankigas “makes white”, blankighas “becomes white”, although there’s no real reason why the syntax should need them; mi blankas la domo is perfectly intelligible as it is. [TYE 178]’s appeals to “international usage” dictating the explicit difference between transitive and intransitive verbs just aren’t convincing; simplicity would be far better here.
Example 3: for no obvious reasons, the syntax of numbers allows the inconsistency of mil bestoj for “a thousand animals” versus miliono da bestoj for “a million animals”. This is because Zamenhof never made up his mind whether or not numbers should be nouns, adjectives or something else. (Thanks to Brendan Linnane again for this one.)
Example 4: “mine” as a pronoun looks like an adjective, la mia “the my”; surely it should be a noun like la mio? Arguably, it could also be la mii, with the pronominal ending, or even la miio.
Example 5: “it (the weather) is warm” is mysteriously translated [TYE 62] as estas varme: what’s wrong with la vetero estas varma? And why (same page) should the idiomatic estas varme al mi be more usual than mi estas varma for “I am warm”? Since the accusative case can replace al, do estas varma mi and estas varme min mean the same thing? If so, this implies that certain combinations of grammatical category and case are equivalent, which is sure to screw up some other parts of the grammar somehwere. This is what happens when you combine ungrammatical idioms with mechanistic principles.
Example 6: Adverbs formed from participles are an unpredictable part of the grammar: how are you supposed to know that “having finished the work, he went home” is fininte la laboron, li hejmeniris [TYE 173]? This is another example of cleverness taking precedence over clarity; something like li finis la laboron kaj iris al hejmo is much clearer. The strange compound word hejmeniris, by the way, is exactly the sort of unwieldy and obscure compound word which Esperantists seem to think is a Good Idea.
Example 7: According to [TYE 158], and for no good reason, you can’t use antaw “before” or post “after” directly before a verb. The constructions you have to use instead are inexplicably different: “I will work before resting” is mi laboros antaw ol ripozi, but “I will rest after working” turns out to be mi ripozos post kiam mi laboros, which suggests Zamenhof couldn’t get past the Latin idiom with post quam.
Example 8: Even an innocuous sentence like mi amas la hundojn is ambiguous; it means both “I like dogs” in a general sense and the more specific “I like the dogs”.
For further contra-propaganda, here are some other pages which criticise Esperanto.
[JBR] Justin Rye’s Learn Not To Speak Esperanto!, written by a linguistics graduate who knows what he’s talking about, is the thoroughly unmissable cannon to this page’s peashooter.
[WIRED] Wired magazine has An interesting article about constructed languages, which makes the encouraging observation that “Esperanto seems to be on the wane”.
[DYER] The Problem of an International Auxiliary Language and its Solution in Ido by L. H. Dyer, 1923, is a long, rather rambling, but ultimately devastating exposition of the defects of Esperanto, in particular the vocabulary.
[JESP] AN INTERNATIONAL LANGUAGE (1928), by the late Otto Jespersen, one of the greatest linguists of the 20th century. Largely a description of Jespersen’s own IAL Novial; contains information about some of the defects with Esperanto.
Part of an essay about a reformed English called Lango. The page contains some material about Esperanto and its faults.
“A web geek’s guide to artificial languages” also has a page about Esperanto and some about other IALs. The writer is critical, in a good-humoured way, of all of the languages, and even links to here!
Other pages of interest
[DH] Don Harlow’s Esperanto site has a lot of information about Esperanto, including a detailed history of the more successful auxlangs.
Rick Harrison’s Artificial Languages Lab contains some interesting pages, such as a summary of a discussion [RHD], which contains some choice quotes and provocative anecdotal evidence, about the merits or otherwise of Esperanto’s vocabulary. Also of note is his set of guidelines for an optimal IAL [RHG]; many of the points are pertinent – in particular the conclusion – and those who consider Esperanto’s design to be perfect or optimal should take note.
Bruce Gilson has some pages with debates about auxlangs which cast further doubt on Esperanto’s claims to greatness.
Mark Rosenfelder’s Language Construction Kit has a short list of some of the irregularites of Esperanto.
Here’s an archive of the conlang mailing list which contains a debate about how complicated Esperanto actually is (it’s a bit over halfway down). The feeling is that it does indeed seem to be more complicated than many Esperantists would like to admit.
If you want a view from the opposite side of the fence, Ken Caviness’s so-called Esperanto FAQ is actually a rather thinly disguised piece of PR in favour of Esperanto, and one of the original inspirations for this critique. As is typical of a lot of Esperanto apologetics, it goes to unnecessary lengths to justify those features of Esperanto which are beneath contempt and can’t avoid occasionally patronizing the reader. Can’t anyone who supports Esperanto discuss it without seemingly trying to put people off?
[PR] Psychological Reactions to Esperanto (here as translated by William Auld). It seems to say that anyone who claims not to like Esperanto is probably not being honest, rather than actually having good reasons for it.
Rick Morneau has some interesting essays about artificial language design.
[TYE] Teach Yourself Esperanto is often recommended as a good book to learn the language from; some sources claim that it’s a sort of definitive guide, even though as we’ve seen it conflicts with other sources in certain details. After three years of searching for a copy, I finally found one in a local charity shop; in a city in which self-teaching texts for languages as little-used as Albanian are routinely on sale in bookshops, this is a remarkable indication of the lack of demand for Esperanto. The book is roughly the same length as many other books in the series, with about the same amount of material, for which reason it’s hard to see how Esperanto can be described as “much easier to learn than any natural language” (back cover). Upon looking through it, two things struck me in particular: how plain ugly Esperanto actually is in any quantity, and how much more besides the 16 rules of grammar you actually need to know.
The Loom of Language by Frederick Bodmer has a very interesting chapter about artificial languages, going into detail about many earlier projects, and is of course healthily critical of Esperanto.