Dear Jon,

I clearly think that your remarks are very relevant to the intention of my question and raise important points on its discussion.

First, I am deeply interested in emergent effects occurring in evolution as I worked on the field of population genetics. If you are interested you can read the joint paper written with Atam Vetta against behavioural genetics, which tries to explain the social behaviour of individuals by the direct action of their own genes. I will be very happy to have your comments on this paper which raises important points on the action of the two considered levels.

Second, I also agree that such concepts as “species” or “group selection” may be very treacherous intellectually. Even for Darwin the term “species” was treacherous when he wrote: “I look at the term species, as one arbitrarily given for the sake of convenience to a set of individuals closely resembling each other, and that it does not differ from the term variety, which is given to less distinct and more fluctuating forms. The term variety, again, in comparison with mere individual differences, is also applied arbitrarily, and for mere convenience sake”. We can only define the descent of a given individual, but I will show in the next paragraph that even an individual is difficult to define.

I know the paradox of the “ship of Theseus”, which will be similar to the paradox of our individual body, which has all its molecules replaced and however remains the same. So that even the individual may be difficult to define.

Do you think that the notion of a mechanism which “consists of entities and activities organized in such a way that they are responsible for the phenomenon” (Illari and Williamson, in European Journal for philosophy of science, 2012, 2, pp 119-135) may lead us to a solution to these problems? Such a mechanism has a material and a formal structure. The material structure accessible through observation or experience may be always changing, while the formal structure constitutes the principles upon which it rests. Such a dual structure may lead us, in my opinion, to a better explanation of what we observe. What is your opinion on this approach?

Are you looking at a kind of fractal effect modeling, with same pattern at micro and marco level ?

I am looking at multilevel modelling, which permits to show simultaneously an individual effect and an aggregate effect of a given characteristic: for example to be a farmer and to live in an area where the proportion of farmers is low or on the contrary high, may lead to a different migratory pattern. See for more details the following chapter of my book on Probability and social science.

Dear Daniel,

If you require to show simultaneously an individual effect and an aggregate effect of a characteristic, the first thing that you would undoubtedly require is identical individuals to perform identical characteristics (in the same time frame if time is also considered as one of the variables)... In the natural world, we know this is not possible...

If however you were to multiply the individual (say through cloning) in order to conduct the study, you would then have this organism living with copies of itself (if the farmer was in the community of farmers as you mentioned)... The observations in this case would eventually not serve the purpose of the study...

Now looking into the modelling aspect of the question, if the model were to portray simultaneous effects of natural systems, the characteristics of the natural world need to be taken into consideration as far as possible, possibly by allowing a margin for deviation from the normal (based on changes in surroundings, time, etc...)...

If on the other hand you were looking into the modelling of inanimate objects that can be made identical, the margin for deviation would reduce and you would have a trouble-free model to deal with...

I hope that this is of any help to you...

Best wishes,

Dominique.

Dear Jon,

I clearly think that your remarks are very relevant to the intention of my question and raise important points on its discussion.

First, I am deeply interested in emergent effects occurring in evolution as I worked on the field of population genetics. If you are interested you can read the joint paper written with Atam Vetta against behavioural genetics, which tries to explain the social behaviour of individuals by the direct action of their own genes. I will be very happy to have your comments on this paper which raises important points on the action of the two considered levels.

Second, I also agree that such concepts as “species” or “group selection” may be very treacherous intellectually. Even for Darwin the term “species” was treacherous when he wrote: “I look at the term species, as one arbitrarily given for the sake of convenience to a set of individuals closely resembling each other, and that it does not differ from the term variety, which is given to less distinct and more fluctuating forms. The term variety, again, in comparison with mere individual differences, is also applied arbitrarily, and for mere convenience sake”. We can only define the descent of a given individual, but I will show in the next paragraph that even an individual is difficult to define.

I know the paradox of the “ship of Theseus”, which will be similar to the paradox of our individual body, which has all its molecules replaced and however remains the same. So that even the individual may be difficult to define.

Do you think that the notion of a mechanism which “consists of entities and activities organized in such a way that they are responsible for the phenomenon” (Illari and Williamson, in European Journal for philosophy of science, 2012, 2, pp 119-135) may lead us to a solution to these problems? Such a mechanism has a material and a formal structure. The material structure accessible through observation or experience may be always changing, while the formal structure constitutes the principles upon which it rests. Such a dual structure may lead us, in my opinion, to a better explanation of what we observe. What is your opinion on this approach?

Dear Dominique,

I understand clearly your objections, but in this case no scientific approach of the individual or the society would be possible. If I follow you, only an objective hermeneutic approach may permit to gain knowledge in the social sciences, and I will not follow this way.

As Bacon said in 1620: “There are and can be only two ways of searching into and discovering truth. The one flies from the senses and particulars to the most general axioms, and from these principles, the truth of which it takes for settled and immovable, proceeds to judgement and to the discovery of middle axioms. And this way is now in fashion. The other derives from the senses and particulars, rising by a gradual and unbroken ascent, so that it arrives at the most general axioms at last of all. This is the true way, but as yet untried”. The first way can be assimilated in great part to what we call now an objective hermeneutic way of thinking, leading to an “interpretative model of the natural, as well as the human, sciences” (Skinner, 1975). The second way is what we call now a scientific one.

But the object of scientific investigation is not phenomena in all their complexity. We will have to retain only those aspects of phenomena that can be characterized by a small number of parameters. In studying migration, I retained, in the example given in this chapter, a number of individual and societal characteristics (in the given chapter they are not all shown, but their effect was not significant), in order to be able to give some results. We can introduce in such an analysis an unobserved heterogeneity, but I don’t think that this may solve the problem. However it had been shown that we may be sure about the meaning of the observed effects, even if we do not know whether all the characteristics affecting them have been introduced in the model.

Best whishes,

Daniel

References:

Bacon F. 1620, The new organon.

Skinner Q. 1975, Hermeneutics and the role of history, New Literary History.

Dear Jon,

I am very happy to see that we are on the same wavelength and that you are also interested in philosophy of science. I am working with a philosopher, Robert Franck, from more than 10 years now and we are editors of the Methodos Series (Springer) whose volumes may be of interest for you. You can have a look at: http://www.springer.com/series/6279?changeHeader . I join here a paper we wrote about demography, saying that if the concern of demography is the behaviour of human populations, from individual to society level, it would end up embracing every thing, which may lead to a hermeneutic approach. We thought that we must strive to focus our research on the more specific object of this science. We are now preparing a more complete paper on this focus.

About Jon Williamson you can have a lot of his papers at the following website: http://www.kent.ac.uk/secl/philosophy/jw/. I think that the paper he wrote with Illari is going on the right way in order to have a mechanistic theory of protein synthesis and natural selection. However, I don’t agree with him when he says that a counterfactual theory and a mechanistic theory may have to be used together.

Finally, I entirely agree with you when you said that you cannot understand people who play golf. It is not only a waste of time but also, in our French Cote d’Azur, a waste of water which is becoming so rare and precious.

With my best regards,

Daniel

This is a great question; thank you. It sounds like you’re asking if one variable in a contained situation has an effect on another, will the effect be consistent across all levels of analysis. The context I like to use is If a person owns a small, one person business and wants to increase sales, he or she can contact various people to get their opinion. Assume that this person closes the business and goes to work for a large organization and is given the task of increasing sales. Committees and teams are common and each individual can affect, not only the solution to the problem, but can affect each other.

When I see questions like this or on similar topics, I go back to the basic principles of Systems Thinking outlined by Banathy and many others. The most relevant for this purpose is that the interaction of the individuals is what drives the system (in this case, the organization); this creates new properties that are specific to the system and are not caused by any one part by itself. This means that through the interactions, what results could not have happened by any one individual alone.

In terms of individual versus multilevel analysis, if a system approach is used, the effect one variable has an effect on another, will not be consistent across all levels of analysis.

Dear Steve,

Thank you very much for your very interesting comments. I agree with you that a system approach may lead to different effects of a variable across the levels of analysis. Because of interactions between agents at a lower level, the behaviour of the aggregate is no more simply the sum of the agents’ action. This non linearity, according to the level, may explain the difference between them.

However it seems necessary for me to go further, in order to find the mechanism which explains how this non linearity occurs. There are many existing theories and models which try to explain how the interactions between agents at a given level may introduce such a non linearity but, in my opinion, they are far from complete and they introduce some conditions not verified by existing populations. For example the agent-based models pre-suppose rules of behaviour and verify whether these micro-based rules can explain macroscopic regularities (Billari and Prskawetz, 2003, p. 2). However, when you look in more details how these models operate, such an objective seems to be more a wish than a reality. In fact, Holland in his last book (2014, p. 48) clearly said that these agent-based models include little provision for agent conglomerates that provide building blocks and behaviour at higher level of organisation. In fact these models always remain with micro-level rules, with the false hope to find a link with the aggregate model. I think that this link cannot be found with this pseudo macro-micro approach.

References:

Billari, F., Prskawetz, A. (2003). Agent-based computational demography. Physica Verlag: Heidelberg, New York.

Holland, J. (2014). Signals and boundaries. The MIT Press: Cambridge.

I think the need to understand the mechanisms is based on our individual purposes, almost like the difference between basic and applied research; both are needed. I might be oversimplifying this, but inherent in systems thinking is the solution to most simple to complex problems in my areas of interest (developmental disabilities, addiction, spirituality). This is partly due to the interrelationship between the components of the systems in my work.

I like that you want to take this further and find the mechanism which explains the non-linearity. As a start, if you look closer at how a system functions at the micro-level, even very small changes between component parts can affect the outcome at the macro-level. On the other hand, properties at the macro-level can cause changes at the micro-level.

If you are asking how these changes happen, I can only point to the principles and properties of systems. This is because, in practice, I know that the outcome will be a resolution of the problem. However, I have to say I would be very interested in the mechanism(s) which explain how the non-linearity occurs. For me, it’s just not necessary to get the results system approaches bring,

Is your interest more theoretical at this point? If you found an answer, what would you do with the information (just curious)?

My interest in this question is not only theoretical but mainly practical, as I think that demography, like many other social sciences, is not a fully formed science but a science in the making (see the joint paper I wrote with Robert Franck on this topic). It needs a more solid ground, than the paradigms it followed since its beginning with the work of Graunt in 1662, in order to become a fully formed and more pertinent science. We need to define the general conditions, without which the demographic phenomena would not be what they are, or would not take place as they do. This will reinforce the scientific pertinence of demograpy and hence its reliability.

Daniel, you provided a great deal of information which helped  to widen the scope of this discussion. In all you contributed, I kept coming back to one word  and that is "define." The need to, not only define conditions, but terminology so we are all talking about the same thing.  I know it is somewhat simplistic to reduce the discussion to one word, but you already did a great job expanding on it. Your points are also important for scientific integrity. Thank you for the excellent articles as well.

Dear Steve,

I was very interested by your approach and by your pertinent questions, as a psychologist, which permitted me to develop more consistently my first ideas on this topic. It is important to discuss some questions in your own field with some researchers working in another field, in order to see the generality of these questions and to have answers that appear applicable in your own field.

Daniel's recent comment encompasses a message that systems thinkers put into practice daily. What we do has an effect on other people, other disciplines, each other, and our extended environment. It is by being aware of interrelationships and ethical conduct that significant improvement for everyone will be possible. ResearchGate is proving to be a much needed vehicle for this.