Journée d’étude,
Lundi 12 juin 2017,
Université Paris Diderot*
Organisation :
Nadine DE COURTENAY (Université Paris Diderot) et Fabien GREGIS (SPHERE)
Présentation
Programme
Résumés
informations pratiques
PRESENTATION
The classical accounts of physical quantities in the epistemology of measurement are yet to be supplemented by the study of a primary concept of modern metrology, the concept of “measurand”. The term “measurand” was introduced by metrologists in the first edition of the International Vocabulary of Measurement (VIM, 1984), in which it was simply defined as “the particular quantity subject to measurement”. But the concept has become more intricate with the definition introduced by the third, current edition of the VIM, “measurand : the quantity intended to be measured”. This definition states that the target of measurement is not necessarily the realized quantity and that the measurand can therefore differ from the property actually measured. The measurand appears to be an entity of a conceptual nature that has to be determined by a description and that involves the intention of the human agent.
The workshop aims to explore the ramifications of the concept of measurand whose proper understanding is crucial for the comprehension of the measurement process. The talks will analyse the VIM definitions and discuss the relation of the concept of measurand to the concept of physical quantity. They will investigate the intentional character of the measurand, the role played by models in its determination, and examine how it connects to the evaluation of uncertainty. Finally, they will confront the particular problems that emerge when one extends the enquiry to such areas as that of consciousness science.
- 9:15 Introduction
- 9:30 Luca MARI (University Cattaneo, Italy)
An epistemology (and some sketches of a related ontology) of quantities
- 10:30 Eran TAL* (McGill University, Canada)
What are measurands ? Operational versus informational conceptions
- 11:30 – 11:45 Pause
- 11:45 Fabien GREGIS (SPHERE, France)
Could there be non-unique "true values" of measurands ? An account of definitional uncertainty
- 12:45 – 14:00 Déjeuner
- 14:00 Marcel BOUMANS (Utrecht University, Nederland)
Measurands outside the laboratory
- 15:00 Pascal LUDWIG & Matthias MICHEL (Paris-Sorbonne University, France)
The problem of measurement in consciousness science
- 16:00 – 16:30 Pause
- 16:30 – 18:30
Round Table : les participants, Nadine DE COURTENAY (Paris Diderot University, France) et Noël DIMARCQ (Observatoire de Paris, France)
- 19:30 Dîner des participants
* Eran Tal donnera une seconde conférence le 13 juin 2017 à l’Université Paris Diderot, salle Malevitch (483A), dans le cadre du Séminaire d’Histoire et philosophie de la physique, intitulée
Measurement, Computer Simulation and Observational Grounding
- An epistemology (and some sketches of a related ontology) of quantities
Luca Mari (University Cattaneo, Italy)
The International Vocabulary of Metrology (VIM, JCGM 200:2012) clearly identifies quantities as the object of measurement, but then is less clear on the nature of the basic relation that through (for example) measurement connects measurands (like the length La of a given rod a) and quantity values (like 1.2345 m) : is it a reasonable attribution (as the VIM suggests in def 2.1), a determination (as stated by the previous editions of the VIM), a representation (as assumed by the representational theories of measurement), ...? or isn’t it rather an equality, as the customary equation La = 1.2345 m supposes ?
An exploration of this issue leads us to better understanding the crucial role of measurand definition in a measurement process.
- What are measurands ? Operational versus informational conceptions
Eran Tal (McGill University, Canada)
The International Vocabulary of Metrology currently defines ’measurand’ as "quantity intended to be measured" (JCGM 2012, 2.3). This definition raises questions about the proper role of human intentions in measurement. Specifically, is it possible for a procedure to measure a different quantity than the one intended ? I distinguish between two opposing views on this matter. Operational views hold that what is being measured is a manifest relation among concrete objects or events, and that the relation realized need not be identical to the one intended. Informational views take measurement to be the location of objects or events in a theoretical parameter space. Such a space, and the measurands specified on it, determine what the relevant procedure will measure. I critically assess both views and discuss their implications for the epistemology and practice of measurement.
- Could there be non-unique "true values" of measurands ? An account of definitional uncertainty
Fabien Grégis (Université Paris Diderot, France)
Multiple examples from various scientific disciplines illustrate that one cannot conceive of a unique “true value” for most usual measurands. In light of this claim, is it possible to understand the seemingly self-contradictory concept of “non-unique true value” of a measurand ? Drawing on the recent developments about the role of models in measurement, I suggest a possible solution. First, I discuss the notion of definitional uncertainty, that describes a vagueness in the measurand definitions, and that metrologists have introduced in response to this issue. I defend that it cannot be considered as an uncertainty in the ordinary sense, and that it rather relates to the idealization choices made into the models in use. Then, I claim that non-unique true values can be conceptualized, provided one acknowledges their approximate character and the approximate character of the equations in which they are involved.
- Measurands outside the laboratory
Marcel Boumans (Utrecht University, Nederland)
According the representational theory of measurement, the most appropriate rule for measurement is a homomorphism that maps the relations between the relevant features of the measurand into a model. The implicit consequence of the homomorphism requirement is that for the measurement to be reliable, the model needs to be as complete as possible. Completeness means here that the model encompasses all possible influences that may affect the measurand. But outside the laboratory our knowledge about those influences is very much dependent on how far they have been observed, that is, appear in the statistics. These observations can be evaluated in an objective way by the application of the relevant statistical methods and techniques, called Type A evaluation. But statistics is not sufficient to make a model complete. Additional knowledge is required, which is often provided by an expert with skilled and experienced knowledge about the measurand and the methods and techniques to measure it. This additional expert knowledge makes the validation of the representation less objective. In system dynamics, however, several validation tests are used that, combined with specific model designs, could be used to make the assessment of incomplete models more objective. This paper suggests that a gray-box model validated by structure-oriented behavior tests, with its constituent modules validated by behaviour pattern tests, is a sufficiently accurate representation to make the measurements sufficiently reliable. As a result, the measurement model does not have to be a homomorphism of the structural relations describing the measurand.
- The measurement problem in consciousness science
Pascal Ludwig & Matthias Michel (Paris-Sorbonne University, France)
A recent trend in consciousness science has been the recognition of the subjective, or phenomenal, aspect of consciousness. Because of this subjective component, consciousness is notoriously challenging to study. Indeed, how should the measurand be defined in this area ? Does it make sense, for instance, to describe a mental state as conscious and measurable if this state cannot be reported by subjects ? In order to dodge these difficult methodological questions, many researchers have been led to rely on purely report-based experimental paradigms, so as to discover neural correlates of consciousness (NCCs). We will call this research program “the NCCs research program” (NCCRP). In this talk, we will argue that the NCCRP is seriously flawed. Our paper has three main goals : first, describe the measurement problem in consciousness science and argue that this problem led to the emergence of methodological artefacts. Second, provide a critical assessment of the NCCs put forward by one of the most popular theory following the NCCRP, the global neuronal workspace theory. Third, provide the means of dissociating genuine NCCs from methodological artefacts.
Université Paris Diderot, bâtiment Condorcet, salle 646A,
4, rue Elsa Morante, 75013 Paris –
plan d’accès
Métro ligne 14 / Station : Bibliothèque François Mitterrand
Métro ligne 6 / Station : Quai de la Gare
RER C / Station : Bibliothèque François Mitterrand
Bus 64 / Stop : Tolbiac-Bibliothèque François Mitterrand
Bus 62 & 89 / Stop : Bibliothèque Rue Mann
Bus 325 / Stop : Watt.
Dans la même rubrique :
- 21e Colloque de la Société Internationale de Philosophie de la Chimie
- Eléments d’histoire de Mésopotamie ancienne à travers les nombres, les mesures et les calculs
- Histoire des techniques d’exploration et de neurostimulation cérébrales
- The implementation of local mathematical practices into the mathematics curriculum
- ED 400, 3e journée d’étude 2016-2017
- ED 400 : Penser le curriculum
- "L’expérience vécue du patient" : sens et implications pratiques
- ERC SAW, colloque de clôture : Writing histories of ancient mathematics – Reflecting on past practices and opening the future
- Journée de l’axe d’histoire et de philosophie de la médecine
- Colloque international des doctorants en Histoire de la philosophie arabe (ED 5)
- Chaire Alicia-Moreau Workshop : "Bacon and French Baconianism"