Multilingualism in Stateless Societies

Document Sample
Multilingualism in Stateless Societies Powered By Docstoc
					Multilingualism and Country Wealth:
     An Empirical Investigation

               David Laitin
     Political Science Department
           Stanford University
             The Sardana
The problem of linguistic normalization in Catalonia
                           • Can Picasso’s image of
                             jubilant carefree nationalists be
                             correct?
                           • Catalonia is the apotheosis of
                             a bourgeois society
                           • Post-Franco nationalism was
                             led by center-right bankers
                           • Linguistic normalization
                             increased a Catalan’s
                             communicative range by zero
                           • What explains their fervent
                             linguistic nationalism?
                           • What cost have the Catalans
                             paid for their nationalism?
An anti-Picasso Explanation
           (and pro Thomas Schelling)

                  The language tipping game (applied to Catalonia)

                                      S

             Pay-offs
             for
             having
             child                                                   C
             educated
             in the                                                      S
             language             C
             of …
                                                                k


                                100%                                 100%
k=tipping point                 Spanish    % of population           Catalan
        How was “tip” achieved?
           My speculations
• Taking Advantage of the Post-Franco Moment
  – Portrayal of inevitability of passing the tipping point
• The Private Subversion of a Public Good
  – On the 1910 ballot, voters rejected an initiative to
    allow liquor sales in cities, which had been prohibited
    in Oklahoma’s original constitution. It was the first of
    several Prohibition-repeal initiatives. The Oklahoma
    humorist Will Rogers would later say, "Oklahomans
    vote dry as long as they can stagger to the polls.“
• Humiliation
  – Avui and La Vanguardia in the metro
Model’s assumption is reasonable:
 Diversity Correlates with Poverty
     Ethnic Fractionalization and GDP/cap
     Ethnic Fractionalization                   GDP/capita 1999 in
                                                $1,000
     Lowest trecile (<.227)                     6.83
     Middle trecile (>.227 % < .589)            6.36
     Highest trecile (> .589)                   2.09
     Bivariate correlation, -0.3987; p=0.0000



• Ethnic fractionalization (probability of a random match
  within country of someone of a different language/ethnic
  group) correlates negatively with country wealth
• Does language diversity breed poverty?
• How costly is the promotion of linguistic diversity in
  Spain?
Measuring the Success of Nationalism: Ethno-
   Linguistic Fractionalization (ELF)

 (Probability that any two randomly matched people in a
           country are of a different ethnicity)


                                     n
       F  1  p                                  2
                                                   i
                                   i 1
Model’s assumption is reasonable:
 Diversity Correlates with Poverty
     Ethnic Fractionalization and GDP/cap
     Ethnic Fractionalization                   GDP/capita 1999 in
                                                $1,000
     Lowest trecile (<.227)                     6.83
     Middle trecile (>.227 % < .589)            6.36
     Highest trecile (> .589)                   2.09
     Bivariate correlation, -0.3987; p=0.0000



• Ethnic fractionalization (probability of a random match
  within country of someone of a different language/ethnic
  group) correlates negatively with country wealth
• Does language diversity breed poverty?
• How costly is the promotion of linguistic diversity in
  Spain?
   Is Spain paying an economic cost for a
     politically inspired increase in ELF?
• I had long assumed that Spaniards in general
  and Catalans in particular have paid an
  economic cost for normalization, though a cost
  the majority of Catalans were willing to pay.
• My presentation today challenges this
  assumption with new, though still preliminary,
  data, and by examining the supposed causal
  pathways from ELFpoverty.
    Path 1: State Rationalization
• Rationalization (Max Weber)
  – States unify weights and measures; legal
    rules; and currencies for purposes of
    efficiency
  – He might also have added language to this list
    of rationalized institutions
  – Rationalized states induce economic growth
• Let us put this theory to test
   Does stateness linguistic unity?
     a look at the data from all countries in the world


• SIL provides data on the number of indigenous
  languages as mother tongue within each
  country;
• Multilingualism is a measure of the number of
  indigenous languages per square kilometer;
• Stateless societies are those countries that
  received independence after 1830, and are not
  commensurate with ancient monarchies (from
  Finer).
       Multilingualism Measure
                   Bottom/Top Five
+--------------------------------------------+
|              country   number indig-   Multiling-
                         enous languages ualism
|--------------------------------------------|
|              BELARUS          1          0 |
|             N. KOREA          1          0 |
|              URUGUAY          2   .0574056 |
|                 CUBA          2   .0598609 |
|            KOREA, S.          2   .0604025 |
|--------------------------------------------|

|--------------------------------------------|
|             CAMEROON        279   .4309955 |
|              NIGERIA        510   .4544135 |
|            INDONESIA        737   .4572725 |
|            SINGAPORE         21   .4911726 |
|           PAPUA N.G.        820   .5144255 |
     Is “Stateless” Relevant?
           Multilingualism and State Societies
Stateless/State          Observations Multilingualism
Stateless                115             .23
State                    40              .25
The bivariate correlation between state and logged
multilingualism (log number of indigenous languages over log
area) is .06, p=.42, not significant at conventional levels; even the
sign is opposite of what would be expected.

Therefore “stateness” does not  lowered levels of societal
multilingualism.
Does Geography Explain Multilingualism?
            (1) Latitude
            (2) Multilingualism and
                 Climate/Soil


         Bivariate Correlation: Mean Growing Season
         (average across all measuring stations) and
         Multilingualism (from Nettle) is .42, p = 0.0004



Nettle’s Interpretation: the longer the growing season, the lower the risk
of starvation; the lower the risk, the less need to establish close social
relations with neighbors (i.e. and learning their language)
(3) Multilingualism and Mountains
                 • For each 1,000 feet of
                   elevation from highest
                   to lowest point in
                   country, log of
                   multilingualism goes
                   up by .27.
                 • This is a map of an
                   island in Papua-New
                   Guinea, the most
                   linguistically diverse
                   country in the world
     (4) Multilingualism and N/S vs. E/W
         Expansion (Jared Diamond)




Bivariate correlation between multilingualism and ratio = .03, and p=.7;
not at all significant
 Preliminary Findings: Geographic Not
  Political Sources of Multilingualism
Variable                           T-stat
                                   (significance)
Elevation Difference               2.80 (p<.01)
Distance from Equator              -2.00 (p<.05)
Mean Growing Season                2.57 (p < .01)
Ratio N-S/E-W                      .61
Stateness                          .42
                       R-squared                    .34
          Geography also Matters for
             Economic Growth!
Country     Ratio: E-W/N-S   Distance from   Mountainous   GDP per capita
                             Equator         Terrain
Nigeria     .81              10              2.4           .93

Mexico      .59              23              31.2          6.8

Spain       .84              40              21.6          11.5

France      1.05             46              10.5          15.4




• Distance from equator is a perfect fit; percentage
  of mountainous terrain works except for Nigeria
  (where forest and desert are effective
  substitutes). Ratio has no predictive power.
    Is multilingualism itself harmful to
             economic growth?
Country   # of languages      Ethnic              Number of    Log               Plural (% whose    GDP per capita
          spoken by >5%       Fractionalization   Indigenous   Multilingualism   mother tongue is
          of the population                       Languages                      dominant
                                                                                 language)

Nigeria   6                   .87                 510          .45               .29                .93

Mexico    2                   .31                 291          .39               .6                 6.8

Spain     4                   .44                 13           .20               .68                11.5

France    2                   .26                 29           .25               .85                15.3




• GDP goes down with rising ethnic
  fractionalization and number of indigenous
  languages, and goes up with the size of the
  largest language group.
  Can States Reduce the Costs of
            Diversity?
                                 State                  Stateless

% who speak dominant
     language
                                   78                         60

% of viable languages
(spoken by >5%) of all
                                   17                         31
   living languages
 GDP/cap, in $1,000
                                   8.8                       3.5

     In State societies, one language becomes dominant; in stateless
     societies, more languages remain viable. This correlates with
     GDP/cap.
          More Evidence on the State
Country     # of languages      Era of universal    Year entry as a   Is there a single   Are regional       GDP per capita
            spoken by >5%       public education    state in int’l    state official      languages given
            of the population   (early=1; late=4)   system            language            official status?

Nigeria     6                   4                   1960              No                  Yes                .93


Mexico      2                   3                   1831              Yes                 No                 6.8

Spain       4                   2                   1816              Yes                 Yes                11.5


France      2                   2                   1816              Yes                 Yes                15.3




• Ancient kingdom, longevity of state institutions, a national
  educational system, and a policy to rationalize state language are
  associated with higher GDP/cap. But granting regional linguistic
  autonomy is different: a positive and significant relationship
  between state recognition of regional languages and higher
  GDP/cap (p<.04).
       Multilingualism and Poverty:
                Conclusions
• Levels of multilingualism are not different for state and stateless
  societies
• Multilingualism is a function of geography: elevation differences,
  proximity to the equator, mean growing season – but not
  length/width of the country.
• For poverty ELF (negatively) and size of dominant linguistic group
  (positively) are statistically related.
• However, well institutionalized states that create national education
  systems, and have a single official language, are positively
  associated with wealth.
• Preliminary finding in re the Catalan normalization question:
  Granting regional autonomy to small regional groups in those highly
  institutionalized states has macro economic benefits, undermining
  the assumption that went into the tipping game.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:4/20/2011
language:English
pages:22