The Epistemological Implications of Critical Complexity Thinking for Operational Research
Round 1
Reviewer 1 Report
Comments on manuscript “The epistemological implications of complexity thinking for operational research” by H.A. Kruger, A.H. Verhoef and R. Preiser
This manuscript aims to apply the ideas of the late South African philosopher Paul Cilliers to the practice of OR. I am an OR practitioner who is familiar with some of the “problem structuring” approaches of OR including soft OR and the multimethodology approach and have some acquaintance with system dynamics modelling and complex adaptive systems (CAS) modelling developed by the late American computer scientist John Holland and others. I am not a philosopher or social scientist and I found the manuscript very hard to understand, extremely repetitious and apparently ignorant of the way OR is actually applied in practice. Much criticism is made of the reductionism of classical OR with its emphasis on mathematics and optimization, even though the pioneers of OR including Ackoff had a much broader vision of OR which corresponds more closely to the way it is actually done, as described for example by Ulrich [37].
The manuscript purports to be about epistemology, but makes no reference to the enormous literature of action learning and action research. I am not familiar with the work of Cilliers, but after having read this manuscript I am none the wiser. I gather from a review of one of his books (Complexity and Postmodernism, http://jasss.soc.surrey.ac.uk/2/2/review1.html ) that he wrote in a clear and interesting way, but feel that this manuscript does not do justice to his ideas. Papers published in the journal “Systems” should be written in such a way as to be understandable by a broad and multidisciplinary audience. See for example the guidelines for article submission to the journal “Complex Adaptive Systems Modeling” (“Complex Adaptive Systems Modeling: A multidisciplinary roadmap”, by Muaz A Niazi, Complex Adaptive Systems Modeling, 2013, 1:1). This manuscript gives lip service to the multidisciplinary nature of OR, but seems to me to be written by and for philosophers. An attempt is made to relate Cilliers’ ideas to similar theories including Critical Systems Thinking [10] and Critical Systems Heuristics [9] but the writing is so turgid that the reader is left none the wiser. Checkland [12, page 96], for example, gives a clear explanation of how the different varieties of systems thinking are inter-related. The manuscript attempts to explain many of its points by reference to the facility location problem of classical OR which can be solved by means of a linear programming model. I don’t think this is a very good example, and is not typical of much OR work. It is also not explained very well. The review of Mingers and White [5] covers hundreds of OR/MS studies, and puts complexity theories into the context of systems thinking.
Like so much writing by social scientists the manuscript is full of jargon, for example page 14 mentions reductionism (of the ontological, epistemological and methodological varieties), determinism, objectivism, empiricism, instrumentalism and technicism. This kind of writing alienates readers coming from a mathematical background, and probably also those coming from the rules-based variety of complexity thinking (CAS) fraternity. If the paper can be substantially re-written addressing these points it may be publishable in this journal. However I consider it may be more suitable for a philosophy journal.
Author Response
Reviewer 1
We would like to thank Reviewer 1 for the comments. Based on the comments received from Reviewer 1, we have made significant changes to the paper (discarding/changing large sections containing philosophical content and jargon that the reviewer referred to) We believe that this will address most of the reviewer’s concerns. We provided answers/explanations to all comments from the Reviewer and the details of those answers are listed below. We hope that Reviewer 1 will find our changes and explanations satisfactory.
Reviewer 1: I am not a philosopher or social scientist and I found the manuscript very hard to understand, extremely repetitious and apparently ignorant of the way OR is actually applied in practice.
Our Response
We have addressed this concern by taking out Section 5 – “General reductionist epistemology of classical operational research” (see lines 682-756). Section 5 was a more philosophical explanation concerning assumptions of OR and scientific method – by removing this section we hope that it is now easier to understand.
We have also change/discard other text (see lines 131-146, lines 336-369, lines 390-392) Most of these sections in the text contains a more philosophical inclined content. We believe that these changes will address the repetitious nature and will make it easier to understand. We have also added an additional explanation/discussion in lines 618 to 670 to more clearly explain how the classical OR concerns raised by Ackoff and Cilliers’ ten complexity principles are linked together. We believe that this will also increase the understanding of the paper.
We acknowledge the view of the reviewer in terms of we are “ignorant of the way OR is applied”. It is difficult to address this general statement as the reviewer did not provide details. We will gladly make changes if the reviewer is more specific and would point out why we are ignorant
Reviewer 1: Much criticism is made of the reductionism of classical OR with its emphasis on mathematics and optimization
Our Response
We have addressed this comment by removing the section (section 5) that focuses on the reductionist nature of OR. In the remaining text of the paper we still refer (much more briefly) to the reductionist nature of OR. However, we are not criticising the reductionist nature of classical OR. We are only re-iterating facts that have already been said by people such as Ackoff and Churchman. See for example, The future of Operational Research is past, by Ackoff. The aim of re-iterating it (without criticising it), is to link the general reductionist nature of OR to the work of researchers such as Midgley, Ulrich and ultimately to the ideas of Cilliers. Our goal is to show that there are opportunities for OR to “learn” from other ideas and applications – this “learning from others” (methodological pluralism) is a valuable approach (see the work of Midgley and Jackson for example)
Reviewer 1: even though the pioneers of OR including Ackoff had a much broader vision of OR which corresponds more closely to the way it is actually done, as described for example by Ulrich [37].
Our Response
We agree with the reviewer – Ackoff had a broad vision of OR and we explicitly acknowledged it by explaining that Ackoff was one of the first thinkers who suggested a problem structuring method (as answer to his own observation of the reductionist nature of OR) called interactive planning (see lines 217-220 in section 2).
Reviewer 1: The manuscript purports to be about epistemology, but makes no reference to the enormous literature of action learning and action research.
Our Response
The aim of the manuscript is to align and/or link OR epistemology to complexity thinking (as described by Cilliers). We did not intend to write a paper on the epistemology of OR related to other approaches and paradigms such as action learning and action research. These types of learning and research refer mainly to the idea of how to improve a process – our aim is not to improve a specific process but rather align OR with complexity as described by Cilliers. Adding something about action research or action learning would not fit in with the scope of the paper which focusses on Cilliers’ work.
Reviewer 1: I am not familiar with the work of Cilliers, but after having read this manuscript I am none the wiser. I gather from a review of one of his books (Complexity and Postmodernism, http://jasss.soc.surrey.ac.uk/2/2/review1.html) that he wrote in a clear and interesting way, but feel that this manuscript does not do justice to his ideas
Our Response
We agree, the book written by Cilliers is clear and written in an interesting way. We are unsure why the reviewer states that our manuscript did not do justice to Cilliers’ ideas [the reviewer admits that he/she is not familiar with Cilliers’ work and has only read a book review of Cilliers’ work]. We have taken the core idea of Cilliers’ book where complex systems are defined/described using ten aspects (see lines 327-335). We then showed how OR can be described in terms of Cilliers’ ten complex characteristics as a complex system (sections 4.1 to 4.10). We have also now added new text (lines 618-670) to further explain how Cilliers’ ideas are linked to, and aligned with classical OR concerns as raised by Ackoff. Showing how a different discipline can be related to Cilliers’ work, we hoped that it is doing justice to the work of Cilliers. If the reviewer can provide us with more detail on how we should do justice to Cilliers’ ideas, we will gladly incorporate this into the manuscript.
Reviewer 1: Papers published in the journal “Systems” should be written in such a way as to be understandable by a broad and multidisciplinary audience. See for example the guidelines for article submission to the journal “Complex Adaptive Systems Modeling” (“Complex Adaptive Systems Modeling: A multidisciplinary roadmap”, by Muaz A Niazi, Complex Adaptive Systems Modeling, 2013, 1:1).
Our Response
We have addressed this by removing section 5 which may be seen as a section that reads more difficult than the other sections. We also add new text (lines 618-670) and remove/change other text (lines 131-146, lines 336-369, lines 390-392) and hope that these changes will make the manuscript more understandable to a broader audience.
We did not consult the author guidelines of other journals such as the one referred to by the reviewer. If we need to consult the guidelines of other journals we will gladly do that if we can get more information on which journals’ guidelines we need to consult.
Reviewer 1: This manuscript gives lip service to the multidisciplinary nature of OR, but seems to me to be written by and for philosophers.
Our Response
Again, we apologise if there is any misunderstanding. We did not discuss (or pay lip service) to the multidisciplinary nature of OR – we discussed the fact that OR can be described as a complex system in terms of Cilliers’ 10 complexity principles such as large number of elements, non-linearity etc.
Reviewer 1: An attempt is made to relate Cilliers’ ideas to similar theories including Critical Systems Thinking [10] and Critical Systems Heuristics [9] but the writing is so turgid that the reader is left none the wiser.
Our Response
We are unsure about this comment by the reviewer. The reference [10] of Jackson is used to motivate the idea of “learning from others” (methodological pluralism) – this is to try and justify that OR may learn from other ideas and approaches such as complexity (see lines 85-88). The work of Ulrich [9] (and Midgley [7]) was briefly mentioned in section 3 sub-section 3.1 because they had, unbeknownst to Cilliers, similar ideas – this gives another opportunity for methodological pluralism and learning from one another (see lines 259-261). We acknowledge in the paper that there are similarities between Cilliers’ ideas and critical systems thinking and critical systems heuristics (see for example section 3.1) – but nowhere are we trying to relate Cilliers’ ideas to these approaches. We relate Cilliers’ ideas to OR as a discipline (see sections 4.1 to 4.10).
Reviewer 1: Checkland [12, page 96], for example, gives a clear explanation of how the different varieties of systems thinking are inter-related.
Our Response
The aim of our manuscript is not to explain the varieties of systems thinking – the aim is to show that OR can be described as a complex system in terms of the 10 complex characteristics of Cilliers and that this may lead to methodological pluralism where one approach (OR) may learn from another (complexity)
Reviewer 1: The manuscript attempts to explain many of its points by reference to the facility location problem of classical OR which can be solved by means of a linear programming model. I don’t think this is a very good example, and is not typical of much OR work. It is also not explained very well.
Our Response
It is not clear why the reviewer thinks that the facility location model is not a good example as he/she does not provide any explanations/details.
There is no one example that is typical of the majority of OR work. However, one of the main and most powerful techniques in OR is the use of a linear programming model – a typical facility location problem can be solved by such a model. We have also provided a reference (see line 469 and reference [49]) to motivate the use of a facility location problem as a suitable example to illustrate the complexity dynamics of a typical OR application.
It is also unclear why the reviewer says that it is not explained very well. We acknowledge that we can give a lengthy (mathematical) discussion of the model and provide details of the mathematical formulations (for different types of facility location problems); variables used; assumptions etc. This is not the idea – the idea is to mention an example and how it fits in with the context of the arguments. Facility location models are an important and prominent application area in OR. There are a large number of different facility location models and the area almost forms a sub-area within OR. See for example the work of MS Daskin (e.g. the text book Network and Discrete Location) to see the importance and scope of facility location problems.
If the reviewer can provide us with more details on why he/she thinks it is a bad example and why it was not explained well, we will make the necessary changes.
Reviewer 1: The review of Mingers and White [5] covers hundreds of OR/MS studies, and puts complexity theories into the context of systems thinking.
Our Response
We agree with the reviewer and we have explicitly mentioned that in lines 41-44.
Reviewer 1: Like so much writing by social scientists the manuscript is full of jargon, for example page 14 mentions reductionism (of the ontological, epistemological and methodological varieties), determinism, objectivism, empiricism, instrumentalism and technicism. This kind of writing alienates readers coming from a mathematical background, and probably also those coming from the rules-based variety of complexity thinking (CAS) fraternity.
Our Response
We have addressed this concern of the reviewer - all the terms that the reviewer referred to appeared in Section 5. We have now removed section 5 and hope that this will significantly reduce the use of terms that may not be familiar to some readers. We also add new text (lines 618-670) and remove/change other text (lines 131-146, lines 336-369, lines 390-392) and hope that these changes will make the manuscript more understandable to a broader audience.
Reviewer 2 Report
This paper surveys the study on operational research mainly from the philosophical perspective. The writing is too abstract to understand. It is written in the abstract: “the aim of the paper is to highlight that by drawing on a general understanding of complexity theory, classical OR approaches can be enriched and broadened by adopting an epistemology based on the …” However, I did not see any clear evidence to support this argument.
I suggest the authors give some more concrete examples or clear introduction to classic operational research or other important concepts such as systems thinking, complexity theory, ect. With the current writing, I am afraid that this paper is only interested to a very limited audience. If the authors do not want to have such an introduction, I then suggest the authors submitting the paper to journals with emphasizes on philosophy.
Author Response
Reviewer 2
We would like to thank Reviewer 2 for his comments. We have addressed the concerns by taking out a complete section (section 5) as well as make changes to other text. We have provided answers/explanations to all the comments from Reviewer 2 – our responses are listed below. We hope that the reviewer will find our explanations and changes satisfactory.
Reviewer 2: This paper surveys the study on operational research mainly from the philosophical perspective. The writing is too abstract to understand. It is written in the abstract: “the aim of the paper is to highlight that by drawing on a general understanding of complexity theory, classical OR approaches can be enriched and broadened by adopting an epistemology based on the …” However, I did not see any clear evidence to support this argument.
Our Response
We have addressed this comment by adding new text (see lines 618-670) to explicitly show how the ten characteristics are aligned/linked to the classical OR concerns as raised by Ackoff. Furthermore, the entire section 5 (section 6 in the first version) focuses on how the objective was achieved. From line 794 onwards we refer explicitly to the objective (using the same words quoted by the reviewer) and explained that new perspectives gained from the understanding of complexity include emergence, setting of boundaries, lack of complete knowledge, and responsibility. We then discussed these four aspects in more detail in subsections 5.1 to 5.4 (6.1 to 6.4 in first version). Following this, we re-iterate (with examples) in lines 938-948 the benefits for OR by acknowledging the principles of complexity. The complete section 5 (6 in first version), with all its subsections (plus newly added text in lines 618-670), provides evidence of how classical OR can be enriched by drawing on general complexity principles.
Reviewer 2: I suggest the authors give some more concrete examples or clear introduction to classic operational research or other important concepts such as systems thinking, complexity theory, ect. With the current writing, I am afraid that this paper is only interested to a very limited audience. If the authors do not want to have such an introduction, I then suggest the authors submitting the paper to journals with emphasizes on philosophy.
Our Response
We have removed Section 5 on the reductionist epistemology of classical OR that contains a more philosophical description of the assumptions of OR as well as a description of a scientific method. We also add new text (lines 618-670) and remove/change other text (lines 131-146, lines 336-369, lines 390-392) and believe that these changes have significantly reduce the philosophical content and make the manuscript more understandable to a broader audience. The focus of the paper is on methodological pluralism where one approach (OR) can learn from another (complexity) and, with the removal/changes as explained, does not over-emphasize philosophy.
We are unsure about the examples that the reviewer refers to. Section 2 provides a discussion on OR while Section 3 provides a discussion on Complexity. We are unsure what additional examples the reviewer requires as he/she did not give any details. The main idea covered in the paper involves OR and Complexity theory as described by Cilliers – we are unsure why, and what type of systems thinking examples the reviewer wants.
Reviewer 3 Report
I feel a bit biased saying that this is a good paper. I read and wrote my dissertation on this topic. The authors make compelling argument. However, I suggest adding recent references as well.
Author Response
Reviewer 3
We would like to thank Reviewer 3 for the comments. We have added new (more recent) references to the text. We trust that the reviewer will find our changes satisfactory.
Reviewer 3: I feel a bit biased saying that this is a good paper. I read and wrote my dissertation on this topic. The authors make compelling argument. However, I suggest adding recent references as well.
Our Response
We have added 4 new and more recent references in sections 1, 5.2 and 5.4 as the reviewer has suggested.
Round 2
Reviewer 1 Report
No comments.
Reviewer 2 Report
The revision is still not easy for me to understand and I feel that it does not fit well the scope of this journal. I suggest the author submit their paper to a more suitable journal.
