&: 
       kager mcneilage & davis 
 
1991: 
       1991 
 
a: 
       it is very encouraging to see a paper on a fascinating 
       and i have to say this is a big but 
       relative to a speech context 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       a clearer statement of the evaluation methods and results achieved would be required 
       submitting later to a speech or phonetics conference such as eurospeech 
       a rare treat in the computational field 
       a formal comparison of indian languages in terms of their phonotactics 
       too much space is devoted to a naive functionalist variety of 
       my little girl would like a statement like 
       and it ls not at all clear why a lexicon should be a 
       it is simply a relation 
       in other words a pronouncing dictionary in speech jargon 
       there is a fair amount of name 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       the topic is more suitable for a speech technology or phonetics conference 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       it is astonishing that in a submission to an acl conference 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
able: 
       should be able to capture information 
 
about: 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       the paper has nothing at all to say about computation 
 
above: 
       taking the points mentioned above into account 
 
acceptable: 
       which is acceptable 
       the paper could be an acceptable acl paper 
 
acceptance: 
       consequently i cannot recommend acceptance for the acl conference 
 
account: 
       taking the points mentioned above into account 
 
achieved: 
       a clearer statement of the evaluation methods and results achieved would be required 
 
acl: 
       and acl is not one of them 
       consequently i cannot recommend acceptance for the acl conference 
       the paper could be an acceptable acl paper 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       it is astonishing that in a submission to an acl conference 
 
acl?: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
 
actual: 
       but without an actual conclusion 
 
actually: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
affect: 
       these comments do not affect the overall quality of the paper 
 
al: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
algebra: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
algorithm: 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       of the algorithm would have been instantly obvious 
 
algorithmic: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
all: 
       and it ls not at all clear why a lexicon should be a 
       having said all this 
       the paper has nothing at all to say about computation 
 
allows: 
       mbrola allows you to impose 
 
almost: 
       it is almost as astonishing that in view of the underspecification problem 
 
alone: 
       let alone finite state methods or dependency parsing 
 
also: 
       also 
       should also be ignored 
 
am: 
       but i am quite amazed that the authors even thought of submitting it to the acl 
 
amazed: 
       but i am quite amazed that the authors even thought of submitting it to the acl 
 
amount: 
       there is a fair amount of name 
 
an: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       the paper could be an acceptable acl paper 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       it is astonishing that in a submission to an acl conference 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
       but without an actual conclusion 
 
and: 
       and so far not too well described language like tamil 
       and the study is to be commended for its pioneering contribution in this field 
       and i have to say this is a big but 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       and acl is not one of them 
       a clearer statement of the evaluation methods and results achieved would be required 
       and re 
       and offers 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
       and vastly more sophisticated 
       and it ls not at all clear why a lexicon should be a 
       and non 
       i find both the languages treated and the constraint 
       if the computational aspects are presented more fully and clearly 
       and the irrelevant and speculative functional phonetic bits thrown away 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       and the authors have clearly researched the problem reasonably thoroughly 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
       and phonetic transcription 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
       and resubmit on another occasion 
 
another: 
       and resubmit on another occasion 
 
any: 
       without any evidence that these approaches are relevant 
 
application: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
applications: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
applied: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
approach: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
approaches: 
       without any evidence that these approaches are relevant 
       genuine fs approaches 
 
appropriate: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
arabic: 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       the pioneering model of arabic using multi 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
are: 
       but there are plenty of conference outlets for speech technology work 
       are really basic set definitions 
       without any evidence that these approaches are relevant 
       if the computational aspects are presented more fully and clearly 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       what makes me hesitate are the following points 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
argued: 
       it could be argued that the paper would be equally relevant to a speech technology conference 
 
as: 
       submitting later to a speech or phonetics conference such as eurospeech 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       it is almost as astonishing that in view of the underspecification problem 
 
aspects: 
       if the computational aspects are presented more fully and clearly 
 
astonishing: 
       it is astonishing that in a submission to an acl conference 
       it is almost as astonishing that in view of the underspecification problem 
 
at: 
       and it ls not at all clear why a lexicon should be a 
       this should at the very least have been demonstrated 
       the paper has nothing at all to say about computation 
 
attempt: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
authors: 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       and the authors have clearly researched the problem reasonably thoroughly 
       that the authors will revise the paper bearing this critique in mind 
 
avms: 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
away: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
basic: 
       basic definitions 
       are really basic set definitions 
 
basis: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
be: 
       and the study is to be commended for its pioneering contribution in this field 
       a clearer statement of the evaluation methods and results achieved would be required 
       should be able to capture information 
       and it ls not at all clear why a lexicon should be a 
       the paper could be an acceptable acl paper 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       the presentation of the problem leaves something to be desired 
       tape fs transducers should not even be mentioned 
       should also be ignored 
 
bearing: 
       that the authors will revise the paper bearing this critique in mind 
 
because: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
 
been: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       of the algorithm would have been instantly obvious 
       this should at the very least have been demonstrated 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
belz: 
       belz 
 
big: 
       and i have to say this is a big but 
 
bits: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
bjorn: 
       bjorn lindblom 
 
both: 
       i find both the languages treated and the constraint 
 
but: 
       but 
       and i have to say this is a big but 
       but there are plenty of conference outlets for speech technology work 
       but would encourage re 
       but not me 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       but even so the methodology is very elementary 
       but without an actual conclusion 
 
by: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
cannot: 
       consequently i cannot recommend acceptance for the acl conference 
 
capture: 
       should be able to capture information 
 
carsonberndsen: 
 
case: 
       in which case the string 
 
casting: 
       casting doubt on the claim of linear complexity 
 
chunks: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
claim: 
       casting doubt on the claim of linear complexity 
 
clear: 
       and it ls not at all clear why a lexicon should be a 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
clearer: 
       a clearer statement of the evaluation methods and results achieved would be required 
 
clearly: 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       if the computational aspects are presented more fully and clearly 
       and the authors have clearly researched the problem reasonably thoroughly 
 
clitic: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
coarsely: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
code: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
commended: 
       and the study is to be commended for its pioneering contribution in this field 
 
comments: 
       these comments do not affect the overall quality of the paper 
 
comparison: 
       a formal comparison of indian languages in terms of their phonotactics 
 
complexity: 
       complexity 
       casting doubt on the claim of linear complexity 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
computation: 
       the paper has nothing at all to say about computation 
 
computational: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       with discussion of its computational properties 
       a rare treat in the computational field 
       if the computational aspects are presented more fully and clearly 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
concludes: 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
conclusion: 
       but without an actual conclusion 
 
conclusions: 
       theoretical or practical conclusions 
 
conference: 
       but there are plenty of conference outlets for speech technology work 
       consequently i cannot recommend acceptance for the acl conference 
       submitting later to a speech or phonetics conference such as eurospeech 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the topic is more suitable for a speech technology or phonetics conference 
       it is astonishing that in a submission to an acl conference 
 
consequently: 
       consequently i cannot recommend acceptance for the acl conference 
 
considerable: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
constraintbased: 
 
context: 
       relative to a speech context 
 
contexts: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
contribution: 
       and the study is to be commended for its pioneering contribution in this field 
 
could: 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       the paper could be an acceptable acl paper 
       it could be argued that the paper would be equally relevant to a speech technology conference 
 
critique: 
       that the authors will revise the paper bearing this critique in mind 
 
data: 
       data structures 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
davis: 
       kager mcneilage & davis 
 
defined: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
defining: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
definitions: 
       basic definitions 
       are really basic set definitions 
 
demonstrated: 
       this should at the very least have been demonstrated 
 
dependency: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       let alone finite state methods or dependency parsing 
 
described: 
       and so far not too well described language like tamil 
 
desired: 
       the presentation of the problem leaves something to be desired 
 
devoted: 
       too much space is devoted to a naive functionalist variety of 
 
dictionary: 
       in other words a pronouncing dictionary in speech jargon 
 
discussion: 
       with discussion of its computational properties 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
disjointed: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
do: 
       these comments do not affect the overall quality of the paper 
 
doubt: 
       casting doubt on the claim of linear complexity 
 
draw: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
eg: 
 
elementary: 
       but even so the methodology is very elementary 
 
encourage: 
       but would encourage re 
 
encouraging: 
       it is very encouraging to see a paper on a fascinating 
 
enough: 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
equally: 
       it could be argued that the paper would be equally relevant to a speech technology conference 
 
et: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
european: 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
 
eurospeech: 
       submitting later to a speech or phonetics conference such as eurospeech 
 
evaluation: 
       a clearer statement of the evaluation methods and results achieved would be required 
 
even: 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       but even so the methodology is very elementary 
       tape fs transducers should not even be mentioned 
 
evidence: 
       without any evidence that these approaches are relevant 
 
explanation: 
       phonetic explanation 
       proponent of this type of explanation 
 
fair: 
       there is a fair amount of name 
 
far: 
       and so far not too well described language like tamil 
 
fascinating: 
       it is very encouraging to see a paper on a fascinating 
 
fate: 
       focussing on the historical fate of schwa 
 
field: 
       and the study is to be commended for its pioneering contribution in this field 
       a rare treat in the computational field 
 
fifth: 
       fifth 
 
finally: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
find: 
       i find both the languages treated and the constraint 
 
finely: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
finite: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       let alone finite state methods or dependency parsing 
       in particular to the finite state methods track 
 
first: 
       first 
       first 
       first 
 
focussing: 
       focussing on the historical fate of schwa 
 
following: 
       what makes me hesitate are the following points 
 
for: 
       and the study is to be commended for its pioneering contribution in this field 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       but there are plenty of conference outlets for speech technology work 
       consequently i cannot recommend acceptance for the acl conference 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the topic is more suitable for a speech technology or phonetics conference 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
       for instance 
 
foremost: 
       no reference is made to the foremost 
 
form: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
formal: 
       a formal comparison of indian languages in terms of their phonotactics 
 
formalisation: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
formalisms: 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
former: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
formulated: 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
 
formulation: 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
formulations: 
       formulations like 
 
fourth: 
       fourth 
       fourth 
 
frequencies: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
fs: 
       genuine fs approaches 
       tape fs transducers should not even be mentioned 
 
fully: 
       if the computational aspects are presented more fully and clearly 
 
function: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
functional: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
functionalist: 
       too much space is devoted to a naive functionalist variety of 
 
genuine: 
       genuine fs approaches 
 
girl: 
       my little girl would like a statement like 
 
gradually: 
       gradually the number of people speaking the schwa 
 
grained: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
grown: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
had: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
hand: 
       on the other hand 
 
has: 
       the paper has nothing at all to say about computation 
 
have: 
       and i have to say this is a big but 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       of the algorithm would have been instantly obvious 
       this should at the very least have been demonstrated 
       and the authors have clearly researched the problem reasonably thoroughly 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
having: 
       having said all this 
 
help: 
       will help to solve this problem 
 
hesitate: 
       what makes me hesitate are the following points 
 
highly: 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
 
historical: 
       focussing on the historical fate of schwa 
 
hope: 
       i would hope 
 
hoped: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
however: 
       however 
       however 
 
i: 
       and i have to say this is a big but 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       i 
       consequently i cannot recommend acceptance for the acl conference 
       i 
       i find both the languages treated and the constraint 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       i would hope 
 
icassp: 
       icassp 
 
icslp: 
       icslp 
 
idea: 
       based idea very interesting 
       the underlying idea of this paper is very interesting 
 
ie: 
 
if: 
       if the computational aspects are presented more fully and clearly 
 
ignored: 
       should also be ignored 
 
imperfectly: 
       so the promise of the title is unfortunately only imperfectly realised 
 
imply: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
impose: 
       mbrola allows you to impose 
 
in: 
       and the study is to be commended for its pioneering contribution in this field 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       a rare treat in the computational field 
       a formal comparison of indian languages in terms of their phonotactics 
       in other words a pronouncing dictionary in speech jargon 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       in which case the string 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       it is astonishing that in a submission to an acl conference 
       in particular to the finite state methods track 
       it is almost as astonishing that in view of the underspecification problem 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
       that the authors will revise the paper bearing this critique in mind 
 
including: 
       including what 
 
indian: 
       a formal comparison of indian languages in terms of their phonotactics 
 
information: 
       should be able to capture information 
 
inheritance: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
instance: 
       for instance 
 
instantly: 
       of the algorithm would have been instantly obvious 
 
interest: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
interesting: 
       this paper makes interesting reading 
       based idea very interesting 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       the underlying idea of this paper is very interesting 
 
interspeech: 
       interspeech 
 
into: 
       taking the points mentioned above into account 
 
irrelevant: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
is: 
       it is very encouraging to see a paper on a fascinating 
       and the study is to be commended for its pioneering contribution in this field 
       and i have to say this is a big but 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       which is acceptable 
       and acl is not one of them 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       the paper is very uneven 
       too much space is devoted to a naive functionalist variety of 
       no reference is made to the foremost 
       it is simply a relation 
       there is a fair amount of name 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
       the topic is more suitable for a speech technology or phonetics conference 
       but even so the methodology is very elementary 
       the underlying idea of this paper is very interesting 
       it is astonishing that in a submission to an acl conference 
       it is almost as astonishing that in view of the underspecification problem 
       so the promise of the title is unfortunately only imperfectly realised 
 
it: 
       it is very encouraging to see a paper on a fascinating 
       and it ls not at all clear why a lexicon should be a 
       it is simply a relation 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       it is astonishing that in a submission to an acl conference 
       it is almost as astonishing that in view of the underspecification problem 
 
its: 
       and the study is to be commended for its pioneering contribution in this field 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       with discussion of its computational properties 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
jargon: 
       in other words a pronouncing dictionary in speech jargon 
 
kager: 
       kager mcneilage & davis 
 
language: 
       and so far not too well described language like tamil 
 
languages: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       a formal comparison of indian languages in terms of their phonotactics 
       i find both the languages treated and the constraint 
 
later: 
       submitting later to a speech or phonetics conference such as eurospeech 
 
latter: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
least: 
       this should at the very least have been demonstrated 
 
leaves: 
       the presentation of the problem leaves something to be desired 
 
let: 
       let alone finite state methods or dependency parsing 
 
lexicon: 
       and it ls not at all clear why a lexicon should be a 
 
like: 
       and so far not too well described language like tamil 
       formulations like 
       my little girl would like a statement like 
 
liked: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
lindblom: 
       bjorn lindblom 
 
linear: 
       casting doubt on the claim of linear complexity 
 
linguistic: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
linguistics: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
little: 
       my little girl would like a statement like 
 
local: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
logic: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
loses: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
ls: 
       and it ls not at all clear why a lexicon should be a 
 
made: 
       no reference is made to the foremost 
 
main: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
makes: 
       this paper makes interesting reading 
       what makes me hesitate are the following points 
 
mbrola: 
       mbrola allows you to impose 
 
mcneilage: 
       kager mcneilage & davis 
 
me: 
       but not me 
       what makes me hesitate are the following points 
 
mentioned: 
       taking the points mentioned above into account 
       tape fs transducers should not even be mentioned 
 
methodology: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       but even so the methodology is very elementary 
 
methods: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       a clearer statement of the evaluation methods and results achieved would be required 
       let alone finite state methods or dependency parsing 
       in particular to the finite state methods track 
 
mind: 
       that the authors will revise the paper bearing this critique in mind 
 
model: 
       the pioneering model of arabic using multi 
 
more: 
       and vastly more sophisticated 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       if the computational aspects are presented more fully and clearly 
       the topic is more suitable for a speech technology or phonetics conference 
 
morphophonology: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
much: 
       too much space is devoted to a naive functionalist variety of 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
multitape: 
 
my: 
       my little girl would like a statement like 
 
naive: 
       too much space is devoted to a naive functionalist variety of 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
namedropping: 
 
natural: 
       natural 
 
neither: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
 
no: 
       no reference is made to the foremost 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
nondeterminism: 
 
nor: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
 
not: 
       and so far not too well described language like tamil 
       these comments do not affect the overall quality of the paper 
       and acl is not one of them 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       but not me 
       and it ls not at all clear why a lexicon should be a 
       tape fs transducers should not even be mentioned 
 
nothing: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       the paper has nothing at all to say about computation 
 
number: 
       gradually the number of people speaking the schwa 
 
obvious: 
       of the algorithm would have been instantly obvious 
 
occasion: 
       and resubmit on another occasion 
 
of: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       these comments do not affect the overall quality of the paper 
       but there are plenty of conference outlets for speech technology work 
       and acl is not one of them 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       with discussion of its computational properties 
       a clearer statement of the evaluation methods and results achieved would be required 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       a formal comparison of indian languages in terms of their phonotactics 
       focussing on the historical fate of schwa 
       too much space is devoted to a naive functionalist variety of 
       gradually the number of people speaking the schwa 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
       proponent of this type of explanation 
       of words 
       the set of pairs 
       there is a fair amount of name 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       of the algorithm would have been instantly obvious 
       casting doubt on the claim of linear complexity 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       the underlying idea of this paper is very interesting 
       the presentation of the problem leaves something to be desired 
       the pioneering model of arabic using multi 
       it is almost as astonishing that in view of the underspecification problem 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
       so the promise of the title is unfortunately only imperfectly realised 
 
offers: 
       and offers 
 
on: 
       it is very encouraging to see a paper on a fascinating 
       focussing on the historical fate of schwa 
       casting doubt on the claim of linear complexity 
       on the other hand 
       and resubmit on another occasion 
 
one: 
       and acl is not one of them 
 
only: 
       so the promise of the title is unfortunately only imperfectly realised 
 
operations: 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
or: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       submitting later to a speech or phonetics conference such as eurospeech 
       let alone finite state methods or dependency parsing 
       the topic is more suitable for a speech technology or phonetics conference 
       theoretical or practical conclusions 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
ordering: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
orderings: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
orthography: 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
 
other: 
       in other words a pronouncing dictionary in speech jargon 
       on the other hand 
 
out: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
outlets: 
       but there are plenty of conference outlets for speech technology work 
 
outnumbered: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
over: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
overall: 
       these comments do not affect the overall quality of the paper 
 
pairs: 
       the set of pairs 
 
paper: 
       it is very encouraging to see a paper on a fascinating 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       these comments do not affect the overall quality of the paper 
       this paper makes interesting reading 
       the paper is very uneven 
       the paper could be an acceptable acl paper 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the paper has nothing at all to say about computation 
       the underlying idea of this paper is very interesting 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
       that the authors will revise the paper bearing this critique in mind 
 
parsing: 
       let alone finite state methods or dependency parsing 
 
parsing?: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
 
particular: 
       in particular to the finite state methods track 
 
people: 
       gradually the number of people speaking the schwa 
 
percentages: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
phones: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
phonetic: 
       phonetic explanation 
       and the irrelevant and speculative functional phonetic bits thrown away 
       and phonetic transcription 
 
phonetics: 
       submitting later to a speech or phonetics conference such as eurospeech 
       the topic is more suitable for a speech technology or phonetics conference 
 
phonotactic: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
phonotactics: 
       a formal comparison of indian languages in terms of their phonotactics 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
 
pioneering: 
       and the study is to be commended for its pioneering contribution in this field 
       the pioneering model of arabic using multi 
 
plenty: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       but there are plenty of conference outlets for speech technology work 
 
point: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
points: 
       taking the points mentioned above into account 
       what makes me hesitate are the following points 
 
portuguese: 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
 
practical: 
       theoretical or practical conclusions 
 
preferable: 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
presentation: 
       the presentation of the problem leaves something to be desired 
 
presented: 
       if the computational aspects are presented more fully and clearly 
 
presenting: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
problem: 
       will help to solve this problem 
       and the authors have clearly researched the problem reasonably thoroughly 
       the presentation of the problem leaves something to be desired 
       it is almost as astonishing that in view of the underspecification problem 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
processing: 
       processing regimes 
 
programming: 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
prolog: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
promise: 
       so the promise of the title is unfortunately only imperfectly realised 
 
pronouncing: 
       in other words a pronouncing dictionary in speech jargon 
 
pronouns: 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
properties: 
       with discussion of its computational properties 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
property: 
       counting property 
 
proponent: 
       proponent of this type of explanation 
 
prosody: 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
provide: 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
 
provides: 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
 
quality: 
       these comments do not affect the overall quality of the paper 
 
quantitative: 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
 
quite: 
       but i am quite amazed that the authors even thought of submitting it to the acl 
 
ranking: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
rare: 
       a rare treat in the computational field 
 
raw: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
reading: 
       this paper makes interesting reading 
 
realised: 
       so the promise of the title is unfortunately only imperfectly realised 
 
really: 
       are really basic set definitions 
 
reasonably: 
       and the authors have clearly researched the problem reasonably thoroughly 
 
recommend: 
       consequently i cannot recommend acceptance for the acl conference 
 
reference: 
       no reference is made to the foremost 
 
regimes: 
       processing regimes 
 
reinhard: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
relation: 
       it is simply a relation 
 
relative: 
       relative to a speech context 
 
relevance: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
 
relevant: 
       without any evidence that these approaches are relevant 
       it could be argued that the paper would be equally relevant to a speech technology conference 
 
remotely: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
 
replaced: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
represent: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
representation: 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
 
required: 
       a clearer statement of the evaluation methods and results achieved would be required 
 
researched: 
       and the authors have clearly researched the problem reasonably thoroughly 
 
resubmit: 
       and resubmit on another occasion 
 
resubmitting: 
 
results: 
       a clearer statement of the evaluation methods and results achieved would be required 
 
revise: 
       that the authors will revise the paper bearing this critique in mind 
 
rewriting: 
 
rule: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
said: 
       having said all this 
 
say: 
       and i have to say this is a big but 
       the paper has nothing at all to say about computation 
 
schwa: 
       focussing on the historical fate of schwa 
       gradually the number of people speaking the schwa 
 
schwadeleted: 
 
second: 
       second 
       second 
 
see: 
       it is very encouraging to see a paper on a fascinating 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
 
semantics: 
       semantics 
 
set: 
       are really basic set definitions 
       the set of pairs 
 
should: 
       should be able to capture information 
       and it ls not at all clear why a lexicon should be a 
       this should at the very least have been demonstrated 
       tape fs transducers should not even be mentioned 
       should also be ignored 
 
simply: 
       it is simply a relation 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
sixth: 
       sixth 
 
so: 
       and so far not too well described language like tamil 
       but even so the methodology is very elementary 
       so the promise of the title is unfortunately only imperfectly realised 
 
solve: 
       will help to solve this problem 
 
something: 
       the presentation of the problem leaves something to be desired 
 
sophisticated: 
       and vastly more sophisticated 
 
space: 
       too much space is devoted to a naive functionalist variety of 
 
speaking: 
       gradually the number of people speaking the schwa 
 
speculative: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
speech: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       relative to a speech context 
       but there are plenty of conference outlets for speech technology work 
       submitting later to a speech or phonetics conference such as eurospeech 
       in other words a pronouncing dictionary in speech jargon 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       the topic is more suitable for a speech technology or phonetics conference 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
spoke: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
standard: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
state: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       let alone finite state methods or dependency parsing 
       in particular to the finite state methods track 
 
statement: 
       a clearer statement of the evaluation methods and results achieved would be required 
       my little girl would like a statement like 
 
statistical: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
stipulation: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
strangely: 
       strangely 
 
stringlengthcounting: 
 
structures: 
       data structures 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
study: 
       and the study is to be commended for its pioneering contribution in this field 
 
submission: 
       it is astonishing that in a submission to an acl conference 
 
submitted: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
 
submitting: 
       submitting later to a speech or phonetics conference such as eurospeech 
       but i am quite amazed that the authors even thought of submitting it to the acl 
 
such: 
       submitting later to a speech or phonetics conference such as eurospeech 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
 
sufficiently: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
suitable: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the topic is more suitable for a speech technology or phonetics conference 
 
syllable: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
synthesis: 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
system: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
taking: 
       taking the points mentioned above into account 
 
tamil: 
       and so far not too well described language like tamil 
 
technology: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       but there are plenty of conference outlets for speech technology work 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       the topic is more suitable for a speech technology or phonetics conference 
 
terms: 
       a formal comparison of indian languages in terms of their phonotactics 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
 
that: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       without any evidence that these approaches are relevant 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       it is astonishing that in a submission to an acl conference 
       it is almost as astonishing that in view of the underspecification problem 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       that the authors will revise the paper bearing this critique in mind 
 
the: 
       and the study is to be commended for its pioneering contribution in this field 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       these comments do not affect the overall quality of the paper 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       a clearer statement of the evaluation methods and results achieved would be required 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       consequently i cannot recommend acceptance for the acl conference 
       taking the points mentioned above into account 
       a rare treat in the computational field 
       focussing on the historical fate of schwa 
       the paper is very uneven 
       gradually the number of people speaking the schwa 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
       no reference is made to the foremost 
       the 
       the set of pairs 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
       the two 
       in which case the string 
       of the algorithm would have been instantly obvious 
       casting doubt on the claim of linear complexity 
       this should at the very least have been demonstrated 
       i find both the languages treated and the constraint 
       if the computational aspects are presented more fully and clearly 
       and the irrelevant and speculative functional phonetic bits thrown away 
       the paper could be an acceptable acl paper 
       on the other hand 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the paper has nothing at all to say about computation 
       the topic is more suitable for a speech technology or phonetics conference 
       but even so the methodology is very elementary 
       the underlying idea of this paper is very interesting 
       and the authors have clearly researched the problem reasonably thoroughly 
       what makes me hesitate are the following points 
       the presentation of the problem leaves something to be desired 
       in particular to the finite state methods track 
       the pioneering model of arabic using multi 
       it is almost as astonishing that in view of the underspecification problem 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
       so the promise of the title is unfortunately only imperfectly realised 
       that the authors will revise the paper bearing this critique in mind 
 
their: 
       a formal comparison of indian languages in terms of their phonotactics 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
them: 
       and acl is not one of them 
 
theoretical: 
       theoretical or practical conclusions 
 
there: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       but there are plenty of conference outlets for speech technology work 
       there is a fair amount of name 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
these: 
       these comments do not affect the overall quality of the paper 
       without any evidence that these approaches are relevant 
 
they: 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
 
third: 
       third 
       third 
 
this: 
       and the study is to be commended for its pioneering contribution in this field 
       and i have to say this is a big but 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       will help to solve this problem 
       this paper makes interesting reading 
       proponent of this type of explanation 
       this should at the very least have been demonstrated 
       having said all this 
       this paper provides interesting quantitative data about the phonotactics of european portuguese 
       the underlying idea of this paper is very interesting 
       that the authors will revise the paper bearing this critique in mind 
 
thoroughly: 
       and the authors have clearly researched the problem reasonably thoroughly 
 
those: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
though: 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       though 
 
thought: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
       but i am quite amazed that the authors even thought of submitting it to the acl 
 
through: 
       imply that the properties of the approach have not been sufficiently thought through or actually applied 
 
thrown: 
       and the irrelevant and speculative functional phonetic bits thrown away 
 
title: 
       so the promise of the title is unfortunately only imperfectly realised 
 
to: 
       it is very encouraging to see a paper on a fascinating 
       and the study is to be commended for its pioneering contribution in this field 
       and i have to say this is a big but 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       though there is plenty of linguistic interest and the application is of considerable relevance for speech technology applications to local languages 
       relative to a speech context 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       will help to solve this problem 
       should be able to capture information 
       submitting later to a speech or phonetics conference such as eurospeech 
       too much space is devoted to a naive functionalist variety of 
       no reference is made to the foremost 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       but i am quite amazed that the authors even thought of submitting it to the acl 
       the paper has nothing at all to say about computation 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       the presentation of the problem leaves something to be desired 
       it is astonishing that in a submission to an acl conference 
       in particular to the finite state methods track 
       mbrola allows you to impose 
 
too: 
       and so far not too well described language like tamil 
       too much space is devoted to a naive functionalist variety of 
 
topic: 
       because neither the topic nor the methodology of the paper are remotely suitable for an acl conference 
       the topic is more suitable for a speech technology or phonetics conference 
 
total: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
track: 
       in particular to the finite state methods track 
       the paper loses track of the main computational point in its disjointed discussion of clitic pronouns 
 
transcription: 
       and phonetic transcription 
 
transducers: 
       pass algorithm could have been formulated more clearly in terms of finite state transducers 
       tape fs transducers should not even be mentioned 
 
treat: 
       a rare treat in the computational field 
 
treated: 
       i find both the languages treated and the constraint 
 
treatment: 
       the treatment of arabic morphophonology in an inheritance system by reinhard et al 
 
twopass: 
 
type: 
       proponent of this type of explanation 
 
types: 
       the syllable ranking function is simply a stipulation of a total ordering over coarsely defined syllable types 
 
underlying: 
       the underlying idea of this paper is very interesting 
 
underspecification: 
       it is almost as astonishing that in view of the underspecification problem 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
underspecified: 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
 
uneven: 
       the paper is very uneven 
 
unfortunately: 
       so the promise of the title is unfortunately only imperfectly realised 
 
union: 
       union 
 
using: 
       the pioneering model of arabic using multi 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
 
variety: 
       too much space is devoted to a naive functionalist variety of 
 
various: 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
 
vastly: 
       and vastly more sophisticated 
 
vennemann: 
       vennemann 
 
very: 
       it is very encouraging to see a paper on a fascinating 
       the paper is very uneven 
       this should at the very least have been demonstrated 
       based idea very interesting 
       but even so the methodology is very elementary 
       the underlying idea of this paper is very interesting 
       and very appropriate for a highly underspecified representation system such as arabic orthography 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
view: 
       it is almost as astonishing that in view of the underspecification problem 
 
w_gw_p: 
 
was: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
 
well: 
       and so far not too well described language like tamil 
 
what: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       what makes me hesitate are the following points 
       including what 
 
which: 
       which is acceptable 
       in which case the string 
 
who: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
why: 
       why was this paper submitted to the acl? what is its relevance to finite state methods and dependency parsing? there is nothing of computational interest in the paper 
       and it ls not at all clear why a lexicon should be a 
 
will: 
       will help to solve this problem 
       that the authors will revise the paper bearing this critique in mind 
 
with: 
       with discussion of its computational properties 
       frequencies and percentages of phones in various contexts with no attempt to draw interesting statistical 
       i had hoped that computational linguistics had grown out of presenting raw prolog code chunks with no discussion of their algorithmic complexity or the data structures they represent 
       the paper concludes with a very naive discussion of prosody and speech synthesis 
 
without: 
       without any evidence that these approaches are relevant 
       there are enough appropriate formalisms for the underspecification problem without using programming code 
       but without an actual conclusion 
 
word: 
       deleted form of the word outnumbered those who spoke the standard form and finally the former replaced the latter 
 
words: 
       of words 
       in other words a pronouncing dictionary in speech jargon 
 
work: 
       but there are plenty of conference outlets for speech technology work 
 
would: 
       what i would have liked to see is a formalisation of the rule system in an appropriate algebra or logic 
       a clearer statement of the evaluation methods and results achieved would be required 
       but would encourage re 
       my little girl would like a statement like 
       would provide a much more finely grained phonotactic basis for defining appropriate orderings 
       of the algorithm would have been instantly obvious 
       it could be argued that the paper would be equally relevant to a speech technology conference 
       a clear formulation in terms of avms and appropriate operations would have been preferable 
       i would hope 
 
you: 
       mbrola allows you to impose