When Language
Policies Fail:
The Problem of
Harold F. Schiffman
University of Pennsylvania
Stockholm, September 2006
Implementation: the ‘Achilles
Heel’ of Language Policy
 Failure of a language policy to have the
outcomes that language planners wish,
can often be attributed to poor
implementation of the policy.
 Why should this be so? Why not blame
other factors?
Maybe it depends on who
decides on language policy
 Language policy—often set (decreed,
determined, ordained) by amateurs
 Novices at language planning:
 Hand down a few decrees
 Make grandiloquent statements,
promulgations, decrees
 Sit back and expect things to just happen.
The devil is in the details…
 Deciding on concrete steps
 Allocation of financial resources
 Devising timetables for completion,
evaluation, enforcement, and crosschecking
 Taking the 'long view' of the process
(may not outlast the impatience of
politicians seeking quick fixes for a
The French Revolution:
the Decret Barère
 January 27, 1794: Barère, on the floor of the
Convention, denounces languages (other than
French) as enemies of the Republic:
 “Federalism and superstition speak Breton;
emigration and hate for the republic speak
German; counterrevolution speaks Italian, and
fanaticism speaks Basque. Let us destroy
these instruments of damage and error!”
 The decree passes, taking on the name of
Barère, its most outspoken proponent.
 But what does this decree actually DO?
The real name:
The act of l’An II, 8 pluviose
 Ordained the teaching of French in all
areas where French was not in use
 Followed by the law of An II, 2 thermidor:
The Convention nationale imposes the
use of French in the formulation of all
public acts.
 And then what?
What happened next?
 Bilinguals are to be recruited to be
trained as teachers
 An école normale for this purpose is set
up in Paris, and tries to find students.
 But most bilinguals are already busy
doing other things, so
 The école normale fails, and is
In other words…
 The much vaunted laws and decrees
were not implemented.
 Personnel were not found and trained
 No mention of funding …
 Nobody followed through…
 But the myth that the law/decree actually
had an effect was born.
 The rhetoric of Barère is remembered as
if it were the text of the law!
Another example: Sri Lanka
 Sinhala-only law of 1956 disenfranchised
Tamil; after many protests, and a
protracted civil war,
 13th and 16th amendments to the
Constitution of 1978 made Sinhala and
Tamil official, and English the ‘link’
So what’s wrong?
 The main problems are the dearth of qualified staff and lack of
institutional machinery
 A large number of vacancies in the post of Tamil typists and
translators in several government institutions had not been filled
 Government had decided that Secretaries of every Ministry, Heads
of Departments, Chief Secretaries of Provincial Councils, Heads of
Government Corporations and Statutory Bodies would be
designated as Chief Official Languages Implementation
Officers (COLIO) responsible for the due implementation of the
Official Languages Policy, and
 Senior Assistant Secretaries (Administration) of every Ministry and
Deputies (Administration) in every Department, Local Body,
Provincial Council, Government Corporation and Statutory Body
would also be designated as official languages implementation
officers in their central and divisional offices
Add more bureaucracy?
 What about training the Tamil typists and
 What about funding for this training?
 What about a timetable for implementation?
 What about penalties for failure to implement,
and rewards for implementing the policy?
 http://www.sundayobserver.lk/2005/10/02/new25.html
French Canada:
confusing status planning with
corpus planning
 Change in status of French, to give it
status in higher domains than before
 Top-down decision from Gov’t of
 Not something speakers of French
 Province of Québec (+ other provinces)
were supposed to implement this
In Québec: confusion of status
and corpus planning
 Question then arose about what kind of
French should be given the status?
 What was ‘good French’?
 Was Canadian French good French?
 Metropolitan French was ‘better’
(especially in the eyes of the rest of
English Canada)
In English Canada
 Other provinces are supposed to
implement, but they have no incentive to
do so
 Federal Government doesn’t have the
jurisdiction to act in the provinces
 So the other provinces have in general
failed to act (Mackey 1983:198)
 Perhaps only the rise (and prestige) of
(totally unplanned) ‘French immersion’
has counteracted this failure
Implementation failure in Singapore
 The Tamil case: what are the issues?
 Singapore’s languages:
 77% Chinese
 14% Malay
 7% of Indian origin
 Of this, 60% speak Tamil, i.e. 4% of the
 ‘Egalitarian’ policy, but is it really?
 Can egalitarianism exist at this level?
Tamil: additional problems
 Singapore in general: ‘exonormic’ policy
 All languages taught use external norms
 Thus in Tamil: Focus on ‘pure’ Tamil
(rather than communicative skills)
 Use of ‘Indian Tamil’ purist norms rather
than spoken Singapore variety
 Children see little economic value for this
Additional issues
 ‘Mother tongues’ are used for ‘moral education’
to prevent spread of amoral western values
 Science, technology, other subjects are taught
in English
 Students become compartmentalized bilinguals
 English has more economic value, so Tamils’
(and others too) allegiance shifts to English
 Students feel they don’t ‘own’ Tamil—it is the
preserve of purists who are never satisfied
Singapore English (Singlish)
 All Singaporeans acquire Singlish (local
variety of English) before they acquire
‘standard’ English, which is based on
BANA norms
 BANA: British-Australian-North American
 Singaporeans view Singlish as a marker
of Singapore identity
 So the Gov’t of Singapore now wants to
annihilate, ban, extirpate Singlish!
‘Speak Proper English’
 Gov’t of Singapore now attempting to get
Singaporeans to abandon Singlish and
‘speak proper English’
 The policy on Tamil makes Singapore
Tamils want to shift to English, but the
Gov’t of Singapore now trying to ban
 Does the right hand know what the left
hand is doing here?
Failure of Tamil policy
 Policy on Tamil lacks clear goals shared
by all
 Policy has no timetable or schedules
 Policy lacks evaluation and enforcement
 No checking to see if goals are being met
 No evaluation metrics other than exit testing
 No carrots, only sticks
Carrot and Stick?
 No carrot: students see no economic
value to language, feel they don’t own it
 Stick: students must pass Cambridge Olevel exams in ‘mother tongue to get into
National University of Singapore
 Any failures or shortcomings result in a
‘blame game’
Blame Game
 Teachers blame the students for ‘lack of love of
the language’
 Students blame MOE and CDB, see teachers
as people who can’t do anything else
 MOE and CDB blame students and parents
(for not speaking ‘good’ Tamil at home)
 Everybody blames someone else
 Nobody sees their own failures, or attempts to
fix the problem
Internal Criticism
 Internal criticism is not tolerated in Singapore,
so internal critics have to pussyfoot around and
couch criticism in coded terms
 Foreigners can critique things, but are mostly
 Academics such as at National Institute of
Education are aware of problem but are
 Ministry of Education and Curriculum
Development Board live in a world untouched
by reality
Some changes may be in the offing
 Top-down planning may now give way to
consideration of factors ‘from below’
 Crisis in shift to English (by all language
groups) now seen as a problem
(especially by Chinese)
 Language policy is not maintaining the
 More money now being allocated for
empirical research: $30 million
Does anyone anywhere have a
sensible implementation policy?
 The European Union faces many
 Constantly enlarging, adding new
 Policy requires translation from every
language to every other language, e.g.
Maltese to Estonian, Slovenian to Irish
 Is this a workable policy? Can it be
European Parliament supports multilingualism
but 'pragmatic solutions' needed over rising
interpretation costs
 See report from Eurolang:
Brussels, Thursday, 07 September
 "The House also considers that multilingualism is an
expression of the EU's cultural diversity, which must be
preserved, and that, therefore, while the increasing
number of official languages calls for pragmatic
solutions in the preparatory work within the institutions,
multilingualism must be guaranteed to ensure the
legitimacy and diversity of the European Union."
 Total cost of all the linguistic services of the EU
institutions, translation and interpretation combined,
represents only 1 % of the total EU budget.
Is there some ambiguity here?
 Goes on to say that In 2003, the EP
spent €4m on interpretation services
made available but not used due to late
requests or cancellations. MEPs also ask
that last-minute cancellations and lastminute requests be discouraged.
 In other words, use the services of
translators judiciously, and save money
Some Resources…
 EUROPA - Translation DG - Enlargement
- Nine new languages
 EUROPA - Translation and drafting aids Home Page
Official and Minority
What about the cost?
Enlargement plan:
 EUgiles\actionplan2005.htm
 Become an interpreter:
 EUROPA - Translation and drafting aids Home Page
 Plans for expansion, training, recruitment
 Scholarships for potential trainees
 Action plan for other years, too.
 EUgiles\enlargement2004.htm
Can this plan succeed?
 Even with the best of implementation schemes,
we must ask:
 Is a plan that makes 20+ languages ‘equal’ likely to
 Is the monetary cost of this plan worth it? (Will the
funds still be there down the road?)
 Is the original egalitarianism set up for the early
‘common market’ six countries now being extended
ad absurdem?
 What do other multilingual nations do?
EU States
Chart predicting costs for translation
with 20 languages:
Another example:
South Africa, with 11 languages
 Is this workable and realistic?
 Or is this ‘compensatory’ –trying
to make up for past wrongs?
 Some of these languages are
spoken by less than 4% of the
South Africa’s Languages: where
spoken, and by number of
0.5 %
What are some other (more
realistic) possibilities?
 India: Three-language formula: many ‘official’
languages, but one ‘national’ language (Hindi)
and one other link language (English). All
citizens supposed to learn all three.
 Soviet Union: Russian as link language, other
languages had regional rights
 Austro-Hungarian Empire: German and Xish
 The principle of TERRITORIALITY, but with a
minimum number of administrative ‘working’
India’s linguistic states:
EU: why not use the model of
India or Switzerland?
 Q: Why not have 2 or 3 working link
 A: France doesn’t want it.
 If 3 languages: (English, French, German).
Most would choose English or German
 If 2 languages (English and French),
Most would choose English (and already
If given a choice of 2 or 3 languages,
 Not enough would choose French…
How I view language policy:
 Consists of overt elements (‘official,
explicit, top-down, written, de jure’…) and
 Covert elements (‘unofficial, grass-roots,
implicit, unwritten, de facto’)
 Covert policy may be:
 Subversive (Catalan in Franco Spain…)
 Complicit: meant as ‘window dressing’ or
‘face-saving’ devices (English in Nagaland)
Unintended consequences?
 Covert policy may be something unintended—
the seeds of the destruction or failure of the
policy are in the policy, but the policy-makers
don’t know it.
 Example: failure of 19th-century German
language schools in America which were
covertly assimilationist but were unaware of the
consequences of the policy.
Stundenplan for 19th century ‘German’
schools in the US: 50% English, 50%
Covert policy:
 Most of the overt
policy is visible at the
 There may be more
to the covert policy
than meets the eye
 The whole policy is
immersed in the
‘linguistic culture’ of
the polity
Covert policy may also be:
 Cynically subversive: the authorities want it to
fail and have planned for it to fail—by setting
unrealistic goals.
 Nefarious and hypocritical: they deny that they
want it to fail (Shohamy: ‘hidden agendas’)
 They have chosen it on condition that it never
be implemented or that it be guaranteed to fail
 Examples: some ‘bilingual’ programs in the US
which are intended to result in ‘replacive’
bilingualism, rather than ‘additive’
So French policy with regard to
EU language policy is…
 On the surface egalitarian, with all languages
on an equal footing, regardless of the cost
 Covertly designed to keep English at bay, no
matter what the cost
 Deeply, cynically hypocritical, since minority
languages in France, though recognized by the
EU Charter, are in fact not recognized since
France has not ratified the Charter!
Q: Are there any effective
language policies?
 Do any policies result in
the outcomes explicitly
planned for them?
 Does any polity structure
its policy to cover all the
points that are required?
 Do any avoid unintended
 Explicit and realistic
 Adequate Budget
 Timetable and schedules
 Periodic evaluation and
 Rewards for achieving
 Penalties for failure
A: I’m still looking…
 Luxembourg does a pretty good job
 Lëtzebuergesch for early grades
 German then phased in for elementary
 French phased in, to be used for
secondary education
 Everybody’s literate in German, some
know French well
Domains are specified…
 Lëtzebuergesch has specific domains:
home, school, humor, satire, funerary
 German: (covert, but) journalism,
parliament, law, business…
 French: higher formal domains
Example of Triglossia in the
domain of Law:
 ”Ein Ausländer, der einer Verhandlung vor einem
luxemburgischen Polizeigericht beiwohnt, wird aus
dem Staunen nicht herauskommen, vor allem dann
nicht, wenn er irgendwo gelesen hat, dass die
Amtssprache in Luxemburg das Französische sei. Er
wird nämlich feststellen, dass die Verhandlungen
ausschliesslich auf Lëtzebuergesch geführt werden.
Der Vertreter der Staatsanwaltschaft und der
Verteidiger aber sprechen beim Requisitorium und
Plädoyer französisch. Wüsste er, dass das schriftliche
Urteil in deutscher Sprache verfasst wird, wäre er
vollends aus dem Konzept gebracht.”
 `A foreigner who observes the proceedings of a
Luxembourg police court will be profoundly
shocked, especially if he has read somewhere
that the official language in Luxembourg is
French. He will note that the proceedings are
conducted exclusively in Lëtzebuergesch . For
the requisitorium and the plea, however, the
Counsels for the prosecution and for the
defense speak French. To then learn that the
written sentencing is done in German would
leave him completely confused.' (Hoffmann
 Administrative
units and
position with
regard to its
Lëtzebuergesch for primary literacy:
In Letzebuergesch, but notice French in the signage
on the wall behind the doctor!
‘He shouldn’t drink any more; he has
to drive.’
I’m sorry; I’ve tried everything you
can afford
The downside:
 Foreign guestworkers have no linguistic rights,
have to learn three other languages
 Can’t become citizens if they can’t pass a test
in Letzebuergesch
 (Europe in general doesn’t do much for guest
worker languages, or non-territorial languages)
 It’s a very small country…
 Language policies may fail if they are too
ambitious, or try to be too egalitarian
 ‘Ambitious’: trying to work with too many
languages, or convert L-varieties to Hvarieties by legislative fiat
 Mostly they fail because they fail to
implement the policy, or because of
hidden agendas.
 An extensive bibliography on the subject
of implementation can be viewed at
 http://ccat.sas.upenn.edu/plc/clpp/bibliogs/implementbiblio.html>
The question of language policy implementation is one that is typically
thought of as problematical in some way--sometimes referred to as the
Achilles Heel of language policy--since the failure of a language policy
to have the outcomes that language planners wish, can often be attributed
to poor implementation of the policy. Frequently, language policy makers
are novices at language planning, and tend to view it as something that
can be, or should be, easily implemented--a few broad strokes to give the
basic outlines of the policy, and one is done. I however tend to see
implementation as the most problematical area of language planning, since
it involves many details--deciding on concrete steps, the allocation of
financial resources, devising timetables for completion, evaluation,
enforcement, and cross-checking--and it may also involve a 'long view' of
the process that may not outlast the impatience of politicians seeking
quick fixes for a problem. This paper will examine language policy
failure in a number of different polities, and try to point out why these
policies have problems, and how the role of implementation is typically
ignored or downplayed.

When Language Policies Fail: The Problem of …