Preliminary Discussion Forum for the 3rd MIC Sorbonne workshop (Paris, Nov. 15-16, 2012)

New Standards for Language Studies
Nouveaux Standards pour les Sciences du Langage

You are not logged in.

Announcement

Announcement:
Researchers (linguists, computer scientists, neurologists, psychologists, sociologists, logicians and philosophers) who are interested in the topics specified in this forum may join us in this preliminary discussion forum by sending their name and surname, affiliation and e-mail address to: You will receive a free subscription account which will allow you to post content.

#1 06/04/12 21:52

Wlodarczyk H
Moderator

Foundations

Due to its origins dating back to investigations on typologically different languages and because it is finely attuned to contemporary research in the field of computer science, the ASMIC theoretical framework exhibits both theoretical and experimental characteristics. An epistemological loop goes from domain observations to theoretical assumptions and back. As a matter of fact, these linguistic efforts started with research on specific linguistic problems. However, in the subsequent period, an intense theoretical work has been done, making it possible to better define such important notions as information and meta-information for the sake of language studies, what finally led to working out on this basis surprisingly simple definitions of such elementary concepts as “predication”, “subject”, “object”, “topic”, “focus” etc. This could be achieved by introducing the psychological notion of centre of attention and by redefining the linguistic notion of old/new information status of utterances.

     Moreover, the ASMIC theory is an attempt to create an integrative theoretical framework for linguistic investigations bringing together concepts and methods from an interdisciplinary background: linguistics (both modern and classical approaches), computer science (especially Artificial Intelligence - AI and Knowledge Discovery in Databases - KDD), logics (mainly classical First Order Logic - FOL), cognitive science, brain neurology and experimental psychology.

      As applied to language studies, the concept of attention centering was first introduced within the framework of computational linguistics (known as Centering Theory, cf Grosz et al. 1986, 1995) with the aim to utilise the discourse coherence for the purpose of natural language processing. Thus the concepts of forward and backward looking centres were coined claiming that information is kept coherent in text while flowing from one utterance to the other. In this theory, the concept of centre of attention concerns only the text (discourse) level and even more importantly, as a phenomenon, attention focusing is not considered as one of the most important motivations of the structure of a single utterance. In other words, the computational Centering Theory did not encompass definitions of the subject and object as governed by the centres of attention on the first meta-informative level (where no contrast is proposed by the speaker between the 'old' and 'new' information), neither were the topic and focus  defined as dually opposed concepts, respectively the topic (as governed by an 'old' meta-informative status contrasting with the 'new' one of the comment) and the focus (as governed by a 'new' meta-informative status contrasting with the 'old' one of the background). However, the Meta-Informative Centering theory is not an alternative to the computational one. Rather, it is a complementary approach; the anaphoric and cataphoric motivations of the meta-informative 'old' and 'new' status of the utterance during the communication act, in the MIC theory approach, matches perfectly well the concepts of backward and forward looking centres in the American Centering Theory, which can therefore be integrated into the MIC framework.

     Although the American theory precedes the European MIC theory in time, each one was created separately. The origins of the MIC theory go back to 1999 when we first used the concept of Centre of Interest (in French “centre d’intérêt”) in two different papers (Wlodarczyk, A. – 1999 and Wlodarczyk, H. - 1999). Since that pioneering time, this French term has been replaced by that of Centre of Attention under the influence of the Grosz-Sidner centering-theoretical framework.

       The proponents  of one of the current mainstream pragmatic theories (namely, the relevance theory) such as Sperber and Wilson (1986) claim that linguistic communication depends more on inferences about speakers’ intentions and representations than on the decoding of language expressions. In the MIC theory, the pragmatic dimension of utterances cannot be modelled without the semantic dimension. In order to interpret a linguistic message,  a correspondence must be established between the linguistic form and its contents, but this cannot be achieved without formal representation devices. The logical background for the formal representation of relations expressed in linguistic expressions consists of  reducing multiple relations to unary and binary relations (Associated Semantics theory, AS). Languages provide speakers with semantic patterns (expressed by verbs and their valencies) that are used to express situations. In addition to such patterns, language requires that speakers choose a point of view on the situation they speak about: among other things, this takes the form of the choice of centres of attention. Meta-informative paraphrases (passive or active, personal or impersonal, topicalised or focalised etc.) are different forms that can be used by the speakers to convey the same informative content. In this regard, human languages are considered as diametrically different from formal languages, hence the latter are an attempt to build a bijective correspondence between form and meaning.

Last edited by Wlodarczyk H (02/05/12 16:38)

Offline

#2 28/06/12 10:38

Zielinska D
Member

Re: Foundations

Replying to prof. Andre Włodarczyk’s statement (private communication)

"Our goal is to open a profound discussion on what is linguistics and what can/should be done in order to bring it nearer to other sciences."

as well as to the introductory part of his post, I would like to make the following comment.

Linguistics from the perspective of empirical sciences.

1.    Empirical sciences aim at constructing and testing explanatory models of objects and processes. In empirical sciences one says that a given phenomenon (its model) has been fully explained (corroborated and tested) when two conditions have been met. First, one has explicated the mechanism which brings about and sustains that phenomenon (object) in terms of some mechanistic laws  and the assumptions made when constructing the given model. Second, the explication proposed implies some hypothesis, which can be and has been tested.
In practice, more often than not, one begins with searching for empirical rules (also called phenomenological laws), which capture patterns in data (the way Kepler did, when he analyzed the data collected by Tycho Brache, finding that the mathematical formula for ellipsis summarizes the observed positions of planets revolving around the Sun). This is where all sorts of Data Mining techniques, etc., come in. Only later does one search for some mechanistic laws, which (along with the assumptions made when constructing the given model) imply the respective formulae – hypothesis. (This was what Newton did in relation to Kepler’s results.)

2.    From the perspective of empirical sciences, only the behaviour of material things can be explained (in the sense of providing its cause). Abstract systems alone are not capable of explaining anything in the sense of providing its cause, but can only offer summation rules - rules summarising typical experience (subsuming under generalization).Therefore, .since abstract systems cannot change by themselves ,when attempting to construct an explanatory model of language from the perspective of empirical sciences language cannot be viewed as an abstract system isolated from its users and use. As Bunge explains:

"Languages do not develop or evolve by themselves and there are no mechanisms of linguistic changes, in particular evolutionary forces. Only concrete things, such as people can develop and evolve. And, of course, as they develop or evolve, they modify, introduce, jettison linguistic expressions. The history of mathematics is parallel: mathematicians do come up with new mathematical ideas, which are adopted or rejected by the mathematical community, but mathematics does not evolve by itself” Bunge (2003: 62)

3.    Consequently, I assume that natural language as a semiotic system is closely (integrally) related to evolving material systems (part of the plastic areas of the brain), and therefore language needs to be treated as a result of a process in which the evolution of form and representation are inseparable and determined by a mechanism dependent on the social function of language. In other words, we argued that language as a semiotic system is inseparable from its function as a tool for communication. Therefore we may repeat after Bunge (2003) that the form and content of language are the integral  parts of the history of specific purposeful interactions between linguistic community members involved in speech acts. During that process certain aspects of participant’s brain state which was evoked as a response to bio-socio-environmental conditions, become correlated with the symbolic elements of language through changes to some plastic areas of the speaker’s brain. So natural language as a semiotic system is a reflection of a dynamic system of individual speech acts produced by human agents located in and interacting with social situations. Grzybek (2006: 12) expressed that idea by saying: “Genesis and evolution of these systems must be attributed to repercussions of communication upon structure.” In other words, an explanation of the existence, properties, and changes of linguistic, (more generally speaking, of a semiotic) system is not possible without treating it as an aspect of the (dynamic) interdependence between structure and function, or in Bunge’s (2003) language, without understanding the mechanism supporting that function of the given semiotic system in a specific environment.

This is so because, as Bunge (ibid.) stresses, the changes in material system components involved in their combining to become a higher level unit are always the result of some input from their environment – outside of the system.  Consequently, in an empirical paradigm, the material system (neural connections) supporting linguistic behaviour of an agent is determined not only by bio-psychological principles (as implied by Chomsky (1986) style view of language), but also by external, socially established principles. In other words, the perspective on language just advocated makes linguistics necessarily an inter-science that straddles biology, psychology, and equally importantly, social sciences. Bunge (2003: 63)

4.    In view of the above, Linguistic explanation is not likely to be possible by means of casual relations. Instead, Altmann (1978) proposes that a likely strategy is looking for a functional explanation similar to ones offered in biology. Therefore, as Altmann continues, it is plausible to assume that language is an aspect of a self-organizing and self-regulating system of members of a linguistic community engaged in communication promoting their co-existence – a special kind of a dynamic system with particular properties brought in line as a result of some sort of economy related to the communicative behaviour of its participants. Or as Grzybek (2006) puts it – “the economic result of communicative processes”.

In the empirical paradigm, the assumption of the self-organizing nature of language implies the fact that the laws modelling language must not have reference to specific objects but be statistical . Another reason for statistical nature of linguistic laws in the empirical paradigm is that since the formation of linguistic objects depends strongly on the history of contingencies and exact data is not available, only statistical hypothesis can be formed. Similarly, Bak (1999: 10) talks about life, (which is also characterized by the variability of its exemplars resulting from its self-organizational character): “A theory of life is likely to be a theory of a process, not a detailed account of utterly accidental details of that process such as the emergence of humans.”

Offline

#3 03/07/12 19:40

Wlodarczyk A
Coordinator

Re: Foundations

In our opinion, we are empiricists. However, we always obey the rule that experiments must be repeated many times before a theory is built, otherwise it would be impossible to compare one theory with another. We are well aware of the dynamic nature of communication, but we distinguish clearly between the system and the process. Signs are discrete entities which are mostly used to express continua. Any dynamic system must have knowledge of these discrete entities in order to be able to undergo any change. For this reason we insist on the (very) partial character of information expressed by language.

Obviously, we accept the constructive (versus innate) character of language acquisition and use. However, in order to make any process run, there must first be a system. Luc Steels et al. showed how the emergence of language can be modelled by computers. The question arises, however, as to what can be done with these results for the practical study of human languages. When asked, Steels replied: "nothing as yet". Nevertheless, this does not mean we should simply wait ...

Our proposal is two-fold:

1) Interactive linguistics - here we integrate all that can be done in Corpus Linguistics, but we want to go further, using the research procedure of data mining. The Corpus Linguistics data need to be manually enhanced by linguists within the framework of KDD if we want to describe languages better (with more precision). We use statistics here in two phases (a) during the corpus search as well as (b) during the data mining. For instance, we have statistical evidence on case morphemes in Polish: the Nominative and the Accusative cases belong to a cluster that is separate from all the other cases. This supports our theoretical position that Subject and Object are main and local centres of attention (or, more properly, Attention Driven Phrases - ADPs). As another example: there are twice as many 'extended utterances' (with Topic and/or Focus) as 'base utterances' (with neither topicalised nor focalised Subject and/or Object).

2) Distributed Grammar - this is a completely new framework in which semantics and pragmatics meet. Syntax is no longer viewed as just a "formal" device. It is given meaning here. In very general terms, the role of syntax is to convey the meta-informative status of communication.

Offline

#4 07/07/12 11:44

Zielinska D
Member

Re: Foundations

1. Ad    "In our opinion, we are empiricists"

I believe there are 2 possible understanding of the notion "empisism". 1st one is common in AI 2 nd one in physical sciences These two notions share the crucial core of being a part of scientific methodology and the choice between them is a matter of the focus of one’s interest. IA is interested in raw data resulting in the so called black box theories, ( to use BUnge’s classification). It could be compared to finding the statistical information in data similarly to what Kepler did when proposing his laws of the movement of planets.  In physics the goal is to find additionally actual mechanistic laws (what Newton did for explaining Kepler’s laws). I believe both approaches are equally valid and “scientific”. Actual mechanisms must account for the observations "mined out{ of the data such as the ones described

    "We use statistics here in two phases (a) during the corpus search as well as (b) during the data mining. For instance, we have statistical evidence on case morphemes in Polish: the Nominative and the Accusative cases belong to a cluster that is separate from all the other cases. This supports our theoretical position that Subject and Object are main and local centres of attention (or, more properly, Attention Driven Phrases - ADPs). As another example: there are twice as many 'extended utterances' (with Topic and/or Focus) as 'base utterances' (with neither topicalised nor focalised Subject and/or Object)."

3. Ad "However, we always obey the rule that experiments must be repeated many times before a theory is built, otherwise it would be impossible to compare one theory with another. "

It is the basic requirement of experimental science that expeiments need to be repeatable. Thus we agree (no need for "however") that the theory needs to be based on statistically derived characteristics of corpuses (or statistically assessable experiments). Some (Grise, Wulf, S, refer to the procedure of representing linguistic concepts with statistical information as operationalization of concepts. If you start from the other end – all sorts of statistical techniques may turn out useful– and certainly data mining pushes forward significantly the choice of statistical methods applied to language .


4. Ad,    "We are well aware of the dynamic nature of communication, but we distinguish clearly between the system and the process. Signs are discrete entities which are mostly used to express continua. Any dynamic system must have knowledge of these discrete entities in order to be able to undergo any change. For this reason we insist on the (very) partial character of information expressed by language.
Obviously, we accept the constructive (versus innate) character of language acquisition and use. However, in order to make any process run, there must first be a system. Luc Steels et al. showed how the emergence of language can be modelled by computers. The question arises, however, as to what can be done with these results for the practical study of human languages. When asked, Steels replied: "nothing as yet". Nevertheless, this does not mean we should simply wait ."

I agree with the necessity of making such an approximation for many practical purposes. However, many believe that strictly speaking the division into syntax and semantix is impossible (Bickard and Campbell, JP 1992, grithith 1994). Trying to accommodate these results gives an alternative research perspective suggesting starting approximating language description with the description of the world around us (possible observable interactions – similar to May’s (2002) pragmemes),  deriving linguistic meaning secondarily and viewing syntax and symbols as respective category names.

It needs to be remembered, however, as pointed out by Brook (1986) that human brain with linguistic capabilities is the result of long evolution. Single cell forms of life arose about 3.5 billion years ago, photosynthetic plants a billion years later, fish about 550 million years ago. Next, at intervals of tens of millions years ago came about insects, reptiles and dinosaurs and mammals. Man appeared on Earth 2.5 million years ago and he invented agriculture less then 20 000 years ago. The oldest evidence that man could write comes from about 5000 years ago and the expert knowledge, when symbolic thinking starts to be needed, has been accumulated for only a couple of hundreds years. If we assume that brain functioning depends on knowledge, how it evolved and how it is organized, it needs to be based on modeling skills once needed for survival such as hearing, moving around in space, interpreting signs indicating location in time. Rodney Brook (1986) concludes that “Such skills, in today’s brain often unconscious, laid foundations for logical reasoning and understanding.” In other words, the functioning of the brain depends on the history of its development, in particular, on specific type of survival challenges faced. As a result, as we well know, the brain is not merely reductionist in its structure and functioning. It consists of units of organization with considerable independence each, (i.e., each of them interacting with different environment) on the one hand, but on the other hand, with non-negligible interdependence between units of adjacent levels, which goes both ways up and down. Note, for instance, that if we cut out part of a tissue, it will stop functioning – it will loose its vital characteristics. In other words, the brain is a system in the sense of empirical sciences (c.f., Bunge 2003) – its every level is built out of units of the lower level, each higher level is characterized by fulfilling a new function, which both depends on and conditions the characteristics of the units of the lower level. The situation with language is parallel.

5. Taking a revese approximation of language allows one, e.g., to model linguistic categorization better than previously – (with algorithms of supervised learning, which assign category membership by establishing which pre-defined category a given element resembles more instead of specifying a set of criteria it needs to meet. (See Zielinska 2007). The same technique is used, e.g.,  in recognizing images in medicine.)

6. Ad. "Our proposal is two-fold:
1) Interactive linguistics - here we integrate all that can be done in Corpus Linguistics, but we want to go further, using the research procedure of data mining. The Corpus Linguistics data need to be manually enhanced by linguists within the framework of KDD if we want to describe languages better (with more precision)."

This cutting-edge research is certainly an outstanding contribution to our knowledge of languge (in methodological agreement with both understanding of the concept of empiricism)

7. Ad
"2) Distributed Grammar - this is a completely new framework in which semantics and pragmatics meet. Syntax is no longer viewed as just a "formal" device. It is given meaning here. In very general terms, the role of syntax is to convey the meta-informative status of communication. Please see MIC Meta-Informative Centering Theory"

I feel we are groping in the same direction by departing from classical syntax-semantics division  - by giving a special role to pragmatics. The major difference being that MIC already considers practical solutions, while I am still in the woods pondering over philosophical issues. If, however, I understand MIC correctly, i.e. that the meta-informative status of communication corresponds in fact to its pragmatic functions, then our eventual perspectives derived by starting from somewhat different initial assumptions will likely coincide on practical grounds. I still need to absorb MIC to a grater extend.

8.
Finally, I would like to make one disclaimer.
When advocating looking at linguistics from the perspective of empirical sciences, I mean it somewhat metaphorically. I do not mean to say that linguistics can be seen as a part of natural science, but that it can utilize its core tools. I mean to say that there is a common nucleus based on logic, mathematics, certain philosophical hypothesis about the nature of the world and its scientific study, derived historically from natural sciences, that all sciences share. What exactly should be treated as the core is, of course, a matter of dispute and philosophical investigations.

Last edited by Zielinska D (07/07/12 13:45)

Offline

#5 18/09/12 09:57

Zielinska D
Member

Re: Foundations

Some additional comments

1.    When assuming a given methodological approximation of language e.g., language as an abstract system in Saussure’s sense ("langue" as opposed to "parole"), it is worth considering, what type of issues can be handled assuming the given approximation. Not every approximation can be used to solve every problem.

2.    When constructing a model in empirical sciences, we need to be able to estimate objectively the values of the parameters postulated the given model assumes, which imposes limitations on what models make sense. The role of such considerations can be illustrated well with Bohr’s model of atom. Bohr suggested a cosmological model of an atom (an electron orbiting a nucleus), whose movement could be modeled with the help of Newton laws. Additionally, he postulated that an electron can orbit a nucleus only on orbits separated from the nucleus by distances, which are in such a relation to each other as successive natural numbers. The problem with his model was, as proven by Heisenberg, that one cannot measure simultaneously the position and momentum of a micro-particle such as an electron and the simultaneous knowledge of which is assumed by the model  (Measurement requires an interaction, e.g. throwing a photon on a given particle and observing a change in its course as the result of the collision. Consequently, measuring, for instance, the momentum of an electron disturbs significantly its position, which, thus, is no longer known.). Attempts at solving that epistemic problem gave impulse for the developments of Quantum Mechanics. One of the solutions proposed was to coin models exclusively in terms of measurable parameters.
 
3.    Linguistic models can initially be fleshed out in a natural language. Yet, following the example just discussed indicates that the concepts used to present a given model need to be at least potentially measurable (expressible in objective terms). One strategy is to express linguistic models, e.g., in terms of the characteristics of the form of a corpus. To illustrate what has just been said, let me offer the following example. It has been proposed that one of the factors affecting the ordering of adjectives in AAN phrases is the degree of relativity of the adjectives employed in it. The arrangement of adjectives with the first adjective tending to be the more gradable one, for reasons too long to quote here, is preferable because it ensures a development of a more efficient semiotic system. The more relative adjective is the one that depends on the scale to which it is applied more, eg., the adjective big depends strongly on the scale of an object it is applied to, while the adjective woolen, much less so. Consider for instance a big planet and a big mouse, a woolen mitten and a woolen carpet. The degree of the relativity of an adjective can be defined objectively (approximated) as the ratio of the number of its tokens in comparative and superlative forms to all its tokens in a sufficiently big corpus.

4.    Another sort of limitations imposed on models of language come from epistemic considerations of meaning. Since there is no generally accessible, ideal meaning, an explanatory mechanism of language use and formation cannot be exclusively reductionist. What community members can share, interactively so, is the ordering of aspects of (relations between) their idiosyncratic meanings. Interactions can be compared objectively because interactive categories are provided in externally objectively comparable way.
        Interactively established categories result in the constitution of individual languages (semiotic systems), which, however, have been sufficiently correlated with each other due to being constituted by a similar set of interactive categories. The other way round, language as a semiotic system impacts interactively accessible categories. Thus, the interactions go both ways. Yet, for methodological purposes, to allow one any sort of predictability, we need to establish either a temporarily stable system of interactive categories (pragmemes) and relate a semiotic system to them, or the other way round, which is a dominant practice. The right choice depends on the purpose of the model being constructed.

Offline

#6 20/09/12 19:13

Wlodarczyk A
Coordinator

Re: Foundations

Dr Danuta Zielinska wrote:

1.    When assuming a given methodological approximation of language e.g., language as an abstract system in Saussure’s sense ("langue" as opposed to "parole"), it is worth considering, what type of issues can be handled assuming the given approximation. Not every approximation can be used to solve every problem.]

I will briefly refer only to the example of the first point of your post entitled "Some additional comments". However, let me also mention that your position concerning the procedural interactive modelling brings evidence to our framework of [Interactive Linguistics].

     Indeed, generalisations often reveal to be inappropriate. This is due - as it is easy to guess - to their abstract and necessarily approximative nature. Since de Saussure's proposal, there were other similar proposals: L. Hjelmslev proposed the distinction System/Process, N. Chomsky - Competence/Performance. The three distinctions overlap in many respects. Myself, I used to talk about the distinction System/Process rather than Langue/Parole. In my vocabulary, however, as I consider that language is an interface to a kind of operating cognitive system, it has at least two views: one of them is static while the other one is dynamic (as it is the case of models of computer operating systems). The static view of language is generally referred to as its "System" while its dynamic view could be named its "Process".

    Therefore, elements of the structure of a linguistic system (langue, competence) and its (de)linearisation operations (functions and services) are distributed in a network of memorised resources (lexicon, valency-based utterance schemata, idiomatic expressions) and concurrent procedures (methods, scripts) while what is known as "language use" (individuation, instantiation of the system) should be viewed as a linguistic communication process (parole, performance). Thus, I claim that the linguistic expressions which are produced by the system are the result of its processes and therefore can be measured.

     Presumably, it might be that this distinction is very close to what F. de Saussure, L. Hjelmslev and N. Chomsky had in mind proposing their distinctions.

Offline

#7 15/10/12 10:21

Zielinska D
Member

Re: Foundations

I fully agree with what prof. Włodarczyk wrote, in particular, with his conclusions:

prof. Włodarczyk wrote:

    Therefore, elements of the structure of a linguistic system (langue, competence) and its (de)linearisation operations (functions and services) are distributed in a network of memorised resources (lexicon, valency-based utterance schemata, idiomatic expressions) and concurrent procedures (methods, scripts) while what is known as "language use" (individuation, instantiation of the system) should be viewed as a linguistic communication process (parole, performance). Thus, I claim that the linguistic expressions which are produced by the system are the result of its processes and therefore can be measured.

In this comment I would like to stress some crucial differences between the historical distinctions mentioned by prof. Włodarczyk (such as competence vs. performance) – to be characterized with statements A below - and between System_w and Process_w. i.e the distinction defined above by prof. Wlodarczyk and characterized with statements B below. (The latter position coincides, I believe, with my view of language presented in "Proceduralny model języka".)

A1. If generative grammars (competence) were to reflect some biological processes in the brain, these processes would need a brain, which is very different from what it actually is and how it actually functions. (Generative grammars implied the existence of very sophisticated grammar genes resulting in brain’s functioning like an automata.) A1 implies the problems indicated in A2 and A3
A2. Chomsky’s competence tacitly assumed that language system could be approximated well enough in terms of sharp Aristotelian categories and crisply defined syntax, which, however, is only a very rough approximation of language.
A3. What generative linguists were supposed to do was to guess algorithms (algebraic formulas for competence) expected to represent types of linguistic expressions These rules generated types, which were to be exemplified by specific sentences – examples of performance. Competence does not explain performance, but merely writes it down. Performance is assessed with the intuition of a linguist. Chomsky’s linguistics, as a purely descriptive discipline, belongs in the Arts.


B1 According to the state of the art in brain research, the brain is a self-organizing and self-regulating organ while language is a self-organizing and self-regulating phenomena strongly dependent on the contingencies of its creation. A likely mechanism of language creation (Process_w) resembles natural selection processes, which implies B2b, and B3
B2: Language elements and structure (Linguistic System_w) are not crisp as revealed by a multitude of its quantitative studies starting with Zipf all the way to Celta’s research, or that presented in the Journal of Quantitative Linguistics, as well as by a bulk of psychological experiments, such as these carried out by Eleonore Rosch.
B3 Objective laws of language can only be statistical in nature, derived from measured statistical characteristics of corpuses. The hypothesis of language being a result of self-organizational and self-regulating activity of the brain ( thus in a sense “an interface to a kind of operating cognitive system”[W] and constituted by that operating bio-psycho-social system) opens also a perspective of looking for bio-social laws explaining why certain phenomenological laws have been found. (It also allows one to overcome the problems with coping with defining basic encodings, which is insolvable in Chomsky’s approach.) What linguists are supposed to do is, first, look for the distribution of patterns in language corpuses to extract laws (phenomenological laws, i.e. language systemw). Second, they should look for the underlying principles in biology, psychology and sociology which caused these laws. Both stages of research can be carried out objectively. Process_w accounts for the System_w. Linguistics is science.

To sum up
Chomsky’s (competence) is nothing more than a system of abstract linguistic categories invented for describing language and its creation belongs in the Arts. Chomsky’s performance is a token of competence tinted with performance errors. It neither influences competence, nor even allows one to evaluate competence objectively with its help. Competence is idealized performance, yet without objective ties to reality. Competence serves to describe performance, not to make any predictions concerning performance.

For Włodarczyk and for myself, the System_w are the statistical laws (distributions) found in language, which can be established objectively and finding them belongs in sciences. The Process_W, is the mechanism, which underlies linguistic production, which, in a given environment, has resulted in the creation of the System_w, and keeps creating (adjusting) it in a way which can be objectively measured, albeit in statistical terms. The outcome of the Process_w, as it turns out, depends on the current state of the System_w and the environment. Therefore Process_w is not an idealized System_w the way competence is idealized performance. Additionally, since language “is an interface to cognitive processes”  (and has been constituted with them) thus the Processw and (consequently) the System_w can eventually be implied by bio-cognitive-social laws, along with the history of its creation.

Offline

#8 16/10/12 23:59

Wlodarczyk A
Coordinator

Re: Foundations

Dr Zielinska D wrote:

For Włodarczyk and for myself, the System_w are the statistical laws (distributions) found in language, which can be established objectively and finding them belongs in sciences. The Process_W, is the mechanism, which underlies linguistic production, which, in a given environment, has resulted in the creation of the System_w, and keeps creating (adjusting) it in a way which can be objectively measured, albeit in statistical terms. The outcome of the Process_w, as it turns out, depends on the current state of the System_w and the environment. Therefore Process_w is not an idealized System_w the way competence is idealized performance. Additionally, since language “is an interface to cognitive processes”  (and has been constituted with them) thus the Processw and (consequently) the System_w can eventually be implied by bio-cognitive-social laws, along with the history of its creation.

Please see my contribution concerning this problem:

Offline

#9 12/11/12 10:35

Zielinska D
Member

Re: Foundations

In reply to professor Włodarczyk’s last post, I would like to stress that I use the term “quantitative linguistics” as a cover term for the methodology relying on strict mathematical tools related to statistics. I did not mean it in opposition, e.g., to corpus linguistics when that data is analized statistically, or data mining based studies. So interactive linguistics, synergetic linguistics, Great Unified Theory, and such - all fall in within the term quantitative linguistics as used by myself. (Once I heard the term Zipfian linguistics used as such general cover term for all such situations mentioned, but I felt that this could be even more confusing for those who have not come across such usage before.)
    For the lack of a better general, cover term, I use the term “quantitative linguistics” even in reference to data extracted from a corpus when combined with statistically analyzed psychological studies - as opposed to research, in which some data from a corpus is interpreted solely intuitively and descriptively by the researcher. I am going to mention such a mixed analysis as an application of my field-based procedural model of language during my second presentation at the III MIC conference this week. I shall be arguing that the model postulated can be employed to assess the utterance meanings of some constructions by considering their usage in corpora and analyzing statistically their possible interpretations provided by a group of study subjects. Such a methodology could even constitute a tool for deciding on the meaning of controversial, publicly uttered sentences in courts. If this methodology were embraced by judicial systems, it would affect the standards present both in media and in political life.

Offline

Board footer

Powered by FluxBB