Académique Documents
Professionnel Documents
Culture Documents
Mikael Parkvall
Stockholm University
The typological similarities between Esperanto and other languages have long
been a matter of debate. Assuming that foreign-language structures are more
easily acquired when they resemble those of the learners native tongue, any can-
didate for a global lingua franca obviously ought to be as typologically neutral as
possible. One common criticism of Esperanto is that it is too European, and thus
less accessible to speakers of non-European languages. In order to provide a more
solid base for such discussions, this paper makes an attempt to quantify the Eu-
rocentricity of Esperanto, employing the features catalogued in the World Atlas of
Language Structures. It is concluded that Esperanto is indeed somewhat European
in character, but considerably less so than the European languages themselves.
Since the birth of Esperanto and the Esperanto movement, there has been a con-
tinuous debate on the advantages and disadvantages of promoting Esperanto as the
main international language. One of the focal points of this discussion has been
the languages typological neutrality, or lack thereof. Sceptics have emphasised that
Esperanto is essentially European in nature, and therefore presumably less easily
learnt by non-Europeans. Proponents, not unexpectedly, have tended to downplay
the similarities between Esperanto and European languages, and instead preferred
to emphasize the logical and exceptionless characteristics of the grammar. The ar-
guments from both sides have essentially rested on gut feeling and the comparison
of bits and pieces of grammar, but to the best of my knowledge, no truly systematic
comparison has ever been made.
The publication of the World Atlas of Language Structures (Haspelmath 2005,
henceforth WALS) allows us to perform precisely such a comparison for the first
time, and the aim of this paper is to quantify the degree to which (if at all) Espe-
ranto is a Euro-centric language.
Language Problems & Language Planning 34:1 (2010), 6379. doi 10.1075/lplp.34.1.04par
issn 02722690 / e-issn 15699889 John Benjamins Publishing Company
64 Mikael Parkvall
The data
Method
Since Esperanto is not featured in WALS at all, the first task was to enter the rel-
evant values for this language into the database. As my own active competence is
somewhat limited, I elicited example sentences from two highly competent speak-
ers of Esperanto,1 who in general provided highly convergent answers. The re-
spondents were explicitly instructed not just to provide possible translations, but
How European is Esperanto? 65
rather the translations they felt would be the most natural counterparts of the
example sentences. This is of some importance since Esperanto (like presumably
most other languages) permits certain alternatives that are only rarely exploited by
speakers. For instance, while it allows for free word order, SVO is by far the most
frequently occurring. The same thing applies to, for example, Russian, which in
WALS is classified as an SVO language, and allowing only the unmarked option
in this study is thus essential in order to ensure comparability. Despite the options
present, Esperanto is therefore treated as SVO in my comparison.
The main problem, however, was not to figure out how Esperanto does this
or that, but rather to understand the exact details behind the existing classifica-
tion. Each WALS chapter contains a few pages of text where each author outlines
their guiding principles, but it is of course difficult to cover all potential problems.
While some features were rather straightforward, others required a considerable
effort to get into the head of the author.2 Some errors no doubt remain, but (with
the exception of Chapters 7778, which were excluded for precisely the reason
that I could not grasp the basis of the classification well enough to apply it with
any confidence), I hope and believe that Esperanto is now relatively correctly rep-
resented.
Comparison
We can now compare the degree of similarity between Esperanto and any other
language or group of languages, provided they are featured in WALS.
I first made a comparison based on individual languages. As already men-
tioned, WALS contains data on about 2,500 languages, but most of these are only
very sporadically featured, and can therefore not be used for comparisons beyond
one feature or a handful of features. I chose to consider only languages with data
for 76 or more of the 137 features.3 This gives us a total of 180 languages with
which to compare Esperanto.
The average language makes the same choices as Esperanto in 44.3% of all
cases, i. e. it performs identically to Esperanto on 61 of the 137 features. The range
is between 32.6% and 69.9%, so, in other words, any given language has between
one third and two thirds of its structure in common with Esperanto.
As can be seen in Table1, the nine languages the most similar to Esperanto are
all Indo-European languages of Europe, and, with the exception of Irish (rank 29),
these are in fact the only such languages of the 180 considered here.
This does indeed convey a picture of Esperanto as a rather Eurocentric
language, although it is interesting to note that ranks 1120 include languages
which are neither Indo-European nor European, such as Brahui, Khasi, Quechua,
66 Mikael Parkvall
Kannada, Drehu, Malagasy and Korean. All of these are more similar to Esperanto
than the average language (44.3%). The ten languages which are the least similar to
Esperanto (32.6%34.7%) are all spoken in Australia or the Americas.
With regard to groups of languages, a few options stand out as particularly wor-
thy of closer examination. First, given the claim that Esperanto is heavily Europe-
an in its structure, we would like to compare it to European languages in general.
However, some European languages are normally considered more European
than others, and the label Standard Average European (SAE) was coined in 1941
by Benjamin Whorf (1956:25) for the core of languages sharing certain key fea-
tures. In the following, SAE is taken to consist of all Germanic and Romance lan-
guages, as well as Western and Southern Slavic and Balkan languages (with Dutch,
German, French and Italian assigned twice their weight; cf. van der Auwera [1998]
and Haspelmath [1998]). Third, Romance languages deserve special attention for
the reason that the majority of the Esperanto lexicon is derived from them (Janton
1993:51), leading to the common laymans view that Esperanto is some kind of
Spanish or very much like Italian. Fourth, we might expect Esperanto to display
strong similarities to languages known by its creator, L.L. Zamenhof. Details vary,
but according to standard biographies, these would seem to have included Pol-
ish, Russian, Yiddish, German, French, Latin, classical Greek, ancient Hebrew and
English.4 Finally, it is of obvious interest to compare Esperanto with the languages
of the world as a whole.
Applying the same technique5 to groups of languages as was just done with
individual varieties, we get the following degrees of similarity:
How European is Esperanto? 67
These figures again suggest that Esperanto is indeed relatively close to those lan-
guages that it could be suspected to resemble, namely European languages in gen-
eral, and in particular those known to its creator.
Some other groups of languages have also been suggested to be remarkably
close to Esperanto. Piron (1981), for instance, argues for a special relationship
between Esperanto and Asian languages. However, my scepticism of his claim
is supported by the fact that the language he especially favours, Mandarin, is in
fact no more similar to Esperanto (44.5%) with regard to the traits examined here
than the average language of the world is. The same author also claims that at the
middle plane [=syntax] Esperanto is indubitably Slavic. While the overall similar-
ity between Esperanto and Slavic is 73.3%, it is 81.5% in the realm of syntax,8 so
Pirons second statement seems indeed to have some merit.
So, according to the method pursued here, Esperanto is indeed a European lan-
guage. But that is not necessarily the end of the story. We could also say that de-
spite a general European-ness, in more than half of all WALS features (74 of 137),
Esperanto has chosen whatever strategy happens to be the most common world-
wide. That is, while the average individual language has a mere 44.3% similarity
with Esperanto, the parameter setting displayed by the artificial language is the
most common one world-wide in 54.0% of the cases.
68 Mikael Parkvall
In this context, one may wonder whether the 54% similarity between Esperanto
and the preferred world-wide option is high or low. It turns out that in this metric,
Daga (a language of New Guinea) is the worlds most representative language, in
that it picks the cross-linguistically most common option in 70.8% of the features
considered.9 At the other end of the scale is the Amazonian language Wari, which
adheres to the standard in only 43.4% of all WALS features. The average language
(both in terms of mean and median) scores 55.4% here, while Esperanto, as just
mentioned, ends up with 54.0% similarities. In other words, Esperanto is slightly
less like the average language, than, as it were, the average language is. Among the
181 languages examined here, Esperanto ranks 110 in terms of agreement with the
worlds most common parameter settings.
Now, with Esperanto ranked 110 out of 181 on the normality list, the (other)
European languages fare far less well. Of the nine Indo-European languages in Ta-
ble1, for instance, all rank lower, in many cases by quite some margin. German, for
instance, is the 178th most normal language (out of 181!), that is, one of the most
exotic there are. We could interpret this as meaning that Esperanto is a slightly Eu-
ropean language, but considerably less so than the European languages themselves.
So, it is in a minority of cases that Esperanto has opted for another strategy
than the most common one world-wide. This is usually but by no means always
one that predominates in Europe.
Table3 sets out the wals features according to whether the value for Espe-
ranto is the preferred one in the world as a whole, or only in Europe, or neither.
Table3. The feature values in Esperanto according to whether these match the preferred
value in the world as a whole, in Europe (but not in the world as a whole), or neither
no. of % wals maps
features
World-wide features 74 54% 1, 2, 5, 6, 7, 8, 9, 10, 11, 13, 15, 16, 18, 19, 20, 24,
26, 28, 33, 34, 35, 37, 39, 40, 42, 43, 45, 48, 51, 52,
55, 58, 59, 60, 61, 63, 64, 65, 66, 68, 69, 73, 74, 79,
80, 82, 83, 84, 88, 90, 91, 93, 94, 96, 100, 104, 105,
106, 108, 109, 111, 112, 114, 116, 117, 120, 124,
125, 128, 129, 130, 131, 135, 137
European features 44 32% 4, 12, 17, 23, 25, 27, 30, 31, 36, 41, 44, 47, 54, 56,
57, 62, 71, 75, 81, 85, 86, 87, 89, 92, 95, 97, 98, 99,
101, 107, 110, 113, 115, 118, 119, 121, 122, 123,
126, 127, 132, 133, 134, 138
Neither European nor 18 13% 3, 14, 21, 22, 29, 32, 38, 46, 49, 50, 53, 67, 70, 72,
world-wide 76, 102, 103, 142
TOTAL 136 100%
How European is Esperanto? 69
Manifestations of Eurocentricity
The tacit assumption thus far, both in this paper and in other writings, has been
that normal is good. The rationale behind this assumption is that a linguistic fea-
ture is taken to be easier to acquire if it matches ones own mother tongue.
The neither European nor world-wide row in Table3 contains features which
are relatively unusual, but this does not necessarily imply that they are dysfunc-
tional in particular in view of Esperantos intended role as an auxiliary second
language.
How European is Esperanto? 71
One person who read an earlier draft of this paper suggested that numbers of
speakers be taken into account. Instead of counting the number of languages pre-
ferring one or another strategy, one could count the number of speakers accus-
tomed (or otherwise) to that strategy. I have chosen not to include such a math-
ematical exercise here, but the result is nevertheless clear, in that it would make
Esperanto look considerably more in line with the global state of affairs. That re-
sult, in turn, could be amplified by counting not only mother-tongue speakers, but
also including second language users. The simple reason for this outcome is that
the European languages are typically spoken by vastly greater communities than
are the native languages of Australia or the Americas.14
This alternative perspective, naturally, would not prove that Esperanto is less
(or more) European than I have portrayed it here, but merely that the worlds pop-
ulation is linguistically more European than the above analysis tends to imply.
Also, since English for the time being the only realistic alternative to Esperanto
as a global lingua franca is the most widely spoken language if L2 speakers are
included, the suggested modus operandi could plausibly lead to English emerging
as a language at least as neutral as Esperanto, in terms of its structural relationship
to languages already known and used by the worlds population. Interesting as
such a comparison might be, it falls beyond the scope of this paper.
Concluding remarks
Though I think it is always useful and desirable to quantify existing gut feelings,
the main conclusion from the above is probably dependent on ones own ideo-
logical position. Esperanto is indeed more European in character than many of its
advocates would have it, but probably less so than many of its opponents would
have predicted (in particular those with only a fleeting acquaintance with its lexi-
con). In other words, in terms of Esperantos structural affinities with the known
range of human linguistic potential, the glass could be seen as either half full or
half empty.
It may be worth bearing in mind that most Esperanto texts in history have
probably been produced by people with a European mother tongue, and, in addi-
tion, this also applies to my two informants. In this context, Lindstedt (2006:48)
points out that Esperanto norms are far more dependent on speakers (as opposed
to active language planning) than most people think. It might therefore be in-
teresting to study the differences between the Esperanto portrayed here and that
used by people without knowledge of a European language. There is an obvious
How European is Esperanto? 73
possibility that such versions would emerge as less Eurocentric. As the language
is created by its speakers, a possible shift in speaker demographics could perhaps
lead to a drift away from European-influenced norms.
Notes
* My thanks go to Pivi Juvonen, Sonja Petrovi Lundberg, Hartmut Traunmller, Joakim En-
wall, Bertil Wennergren and two anonymous reviewers for their constructive input. I also owe
special credit to Hkan Lundberg, whose questions provided the initial impetus for this article.
1. Native speakers of German and Swedish respectively; both use Esperanto as their home lan-
guage with their spouses who are of different linguistic backgrounds.
2. Contacts with some of the WALS contributors provided some help here.
3. Any such cut-off point is bound to be arbitrary, of course. In this case, I picked 76 simply in
order to include Polish, a language of special interest here because it was spoken by Esperantos
creator.
4. WALS does not include extinct languages, and therefore excludes Latin, and only features the
modern versions of Greek and Hebrew. It also has extremely few mentions of Yiddish, which
was one of the languages Zamenhof grew up speaking.
5. I am here using the majority option within each group as the group value.
6. This figure includes the modern versions of Greek and Hebrew as stand-ins for the classical
varieties.
7. It should be borne in mind here that labels such as American languages and Oceanic lan-
guages (as is common practice in linguistic typology) refer to speech varieties indigenous to the
respective continents. While the Americas and Oceania are nowadays completely dominated by
(originally) European languages, these are still counted as European, while only the pre-1492
languages are treated as American.
9. Note that this refers only to the features for which there are any data in WALS. Also, as al-
ready mentioned, I included languages with 76 or more attestations in the 137-feature database.
For Daga, for instance, 89 values are included. It is of course possible that Daga behaves excen-
trically with regard to the 48 features about which WALS is silent.
10. Here and in the following, English morphemes are used only to illustrate the types of con-
structions concerned. For readers with a primarily passive command of English, their ungram-
matical nature is marked by means of an asterisk.
11. One of the anonymous reviewers points out that there is some variability among speakers of
Esperanto here, and at least for many users, po fails to display the characteristuics of a preposi-
tion proper.
74 Mikael Parkvall
12. Note, however, that one of the anonymous peer reviewers has reservations about my ac-
count here. S/he writes (originally in Esperanto) that Instead of k-iu one could also propose
ki-u (compare the coinages alies, aliam, and not e.g. ali-ies: such forms must contain an -i-, but it
is felt to belong to the first, not the second part of the word. Moreover, the reviewer reports on
two native speakers that they misspell ne niu as two words, but never nen iu.
13. According to a colleague who is more familiar with the second language acquisition litera-
ture than I, this claim appears not to have been empirically tested.
14. Even Esperanto itself must be considered a larger-than-average language when L2 speakers
are included.
References
Sammanfattning
ermlet, borde givetvis ett internationellt hjlpsprk vara s typologiskt neutralt som mjligt.
Ett vanligt stycke kritik mot esperanto r att det r alltfr europeiskt, och drmed mindre
tillgngligt fr talare av utomeuropeiska sprk. I syfte att erbjuda en bttre grund fr sdana
diskussioner frsker denna artikel med hjlp av de sprkdrag som listas i World Atlas of
Language Structures kvantifiera den eventuella eurocentriskheten hos esperanto. Slutsatsen
r att esperanto mycket riktigt r en smula europeiskt till sin karaktr, dock i betydligt mindre
utstrckning n de europeiska sprken sjlva.
Resumo
As explained earlier, five chapters were excluded from consideration here. These are bracketed
in the following table. The labels are not necessarily self-explanatory, but it would take up dis-
proportionate amounts of space to define them here. The interested reader is therefore referred
to Wals itself for precise definitions.
1 Consonant Inventories
2 Vowel Quality Inventories
3 Consonant-Vowel Ratio
4 Voicing in Plosives and Fricatives
5 Voicing and Gaps in Plosive Systems
6 Uvular Consonants
7 Glottalized Consonants
8 Lateral Consonants
9 The Velar Nasal
10 Vowel Nasalization
11 Front Rounded Vowels
12 Syllable Structure
13 Tone
14 Fixed Stress Locations
15 Weight-Sensitive Stress
16 Weight Factors in Weight-Sensitive Stress Systems
76 Mikael Parkvall
17 Rhythm Types
18 Absence of Common Consonants
19 Presence of Uncommon Consonants
20 Fusion of Selected Inflectional Formatives
21 Exponence of Selected Inflectional Formatives
22 Inflectional Synthesis of the Verb
23 Locus of Marking in the Clause
24 Locus of Marking in Possessive Noun Phrases
25 Locus of Marking: Whole-language Typology
26 Prefixing vs. Suffixing in Inflectional Morphology
27 Reduplication
28 Case Syncretism
29 Syncretism in Verbal Person/Number Marking
30 Number of Genders
31 Sex-based and Non-sex-based Gender Systems
32 Systems of Gender Assignment
33 Coding of Nominal Plurality
34 Occurrence of Nominal Plurality
35 Plurality in Independent Personal Pronouns
36 The Associative Plural
37 Definite Articles
38 Indefinite Articles
39 Inclusive/Exclusive Distinction in Independent Pronouns
40 Inclusive/Exclusive Distinction in Verbal Inflection
41 Distance Contrasts in Demonstratives
42 Pronominal and Adnominal Demonstratives
43 Third Person Pronouns and Demonstratives
44 Gender Distinctions in Independent Personal Pronouns
45 Politeness Distinctions in Pronouns
46 Indefinite Pronouns
47 Intensifiers and Reflexive Pronouns
48 Person Marking on Adpositions
49 Number of Cases
50 Asymmetrical Case-Marking
51 Position of Case Affixes
52 Comitatives and Instrumentals
53 Ordinal Numerals
54 Distributive Numerals
55 Numeral Classifiers
56 Conjunctions and Universal Quantifiers
57 Position of Pronominal Possessive Affixes
58 Obligatory Possessive Inflection
59 Possessive Classification
60 Genitives, Adjectives and Relative Clauses
61 Adjectives without Nouns
62 Action Nominal Constructions
63 Noun Phrase Conjunction
How European is Esperanto? 77
Authors address
Institutionen fr lingvistik
Stockholms universitet
SE-106 91 Stockholm, Sweden
parkvall@ling.su.se
How European is Esperanto? 79