The historical roots (1880–1950) of recent contributions (2000–2017) to ecological economics: insights from reference publication year spectroscopy

ABSTRACT

Several contributions have examined the field of ecological economics using bibliometric techniques. However, none of those bibliometric studies has yet investigated the historical roots of the recent contributions to the field. The aim of this article is to fill that gap, enabling us to test and shed light on the received view that ecological economics is based on pluralistic and interdisciplinary foundations. To do so, we use citations analysis and reference publication year spectroscopy. We focus on the articles published in the journal Ecological Economics from 2000 to 2017 and study the references cited in these articles published from 1880 to 1950. We first examine the ten most influential cited contributions and identify the ten most influential reference publication years. Our results offer qualifications to the received view, supporting low interdisciplinarity and providing caveats concerning pluralism.

KEYWORDS:

• Bibliometrics
• citation analysis
• ecological economics
• heterodox approaches
• interdisciplinarity
• reference publication year spectroscopy

JEL CLASSIFICATION:

• B20
• Q57
• B41

View correction statement:

Correction

Acknowledgement

This article could not have been written without Lutz Bornmann’s and Andreas Thor’s help, advice, and encouragements. I am very grateful to two anonymous reviewers for their comments and suggestions. This research was first presented at ESSCA School of Management and I thank my colleagues for their feedbacks. I am also indebted to François Claveau, Manuel Jesús Cobo, Guillaume Detchenique, Emmanuelle Dutertre, Cyril Fouillet, Mélanie Gastineau, Gwenaëlle Lairet, Antoine Missemer, Anne Musson, Silvia Secchi, Gabriel Weber, and Marx Werner for comments and suggestions. I am also very grateful to Robert Lysiak for his help with the English of the article. Working paper versions of this research were available under the title “The Historical Roots of Recent Contributions to Ecological Economics: A Note Using Reference Publication Year Spectroscopy” as well as under the current title on the author’s SSRN homepage.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes on contributor

Matthieu Ballandonne is an Assistant Professor in economics at ESSCA School of Management. His research interests include the history and methodology of the economics of science, growth, and innovation. He is also interested in the use of bibliometric tools in the history and methodology of economics.

Notes

1 We use the expression ‘ecological economics’ to refer to ecological economics as a field of research and use EE to refer to the journal Ecological Economics.

2 We define inter- and trans- disciplinarity in Section 5. Pluralism means relying on a variety of theoretical approaches, including ‘heterodox’ ones. The definition of the notion of ‘heterodox’ is very problematic. Indeed, what is ‘heterodox’ or ‘orthodox’ is historically evolving and, even at one point in time, the boundaries between the two notions are not clear-cut. For instance, Hodgson (2007) has argued that evolutionary and institutional approaches can now be considered as the ‘new mainstream’ (see also Colander, 2000). We stick to the notion of ‘heterodox’ for two reasons. First, the notion is widely used by ecological economists (for recent examples, see Pirgmaier, 2017; Spash, 2012, 2013a, 2013b). Second, this note is not intended as a history of, or a methodological perspective on, those notions (see Lee, 2009). We thus need a working definition of ‘heterodox’ economics. We use the definition offered by Spencer (2013) and recently used in an article published in EE by Pirgmaier (2017): heterodox economics is

a variety of approaches including ‘old’ institutional economics, post-Keynesian economics, feminist economics, Austrian economics, and Marxian economics. These diverse and sometimes conflicting perspectives have for many years offered an approach to economics that situates the economy in a broader social, historical, and political context. They have by their nature and often by their design sought to develop links with other human and social sciences and to push economics in a more interdisciplinary direction. (Spencer, 2013, p. 5 quoted in Pirgmaier, 2017, p. 56 n. 11)

3 Recent applications of RPYS to various fields of research include, for instance, Barth, Marx, Bornmann, and Mutz (2014) on the Higgs boson, Bornmann, Haunschild, and Leydesdorff (2018) on Eugene Garfield's publications, Comins and Hussey (2015a, 2015b) on, respectively, the Viterbi algorithm and the global positioning system, Elango, Bornmann, and Kannan (2016) on tribology, Haunschild, Barth, and Marx (2016) on density functional theory, Hou (2017) on citation analysis, Khasseh and Mokhtarpour (2016) on knowledge management, Leydesdorff, Bornmann, Marx, and Milojevic (2014) on imetrics, Marx and Bornmann (2014) on Darwin's finches, Rhaiem and Bornmann (2018) on academic efficiency studies, Wray and Bornmann (2015) on philosophy of science, and Yeung (2017) on functional magnetic resonance imaging studies of taste and food.

4 As we shall emphasize in the conclusion, to the best of our knowledge RPYS has not been applied yet in the methodology and history of economics. Edwards et al. (2018) provide a survey of the contributions that use quantitative (including bibliometric) tools in these two fields. As an example, we can compare RPYS with two recent contributions using bibliometric techniques: Gingras and Schinckus (2012) and Claveau and Gingras (2016). Gingras and Schinckus (2012) study the institutionalization of econophysics and its relationships to other disciplines using different bibliometric indicators. In some instances, they do examine the cited references of their dataset but they focus on the journals or disciplines cited. In other words, they do not examine the ‘historical roots’ of econophysics. Claveau and Gingras use the tools of bibliographic coupling, automated community detection in dynamic networks, and text mining in order to get a ‘macro’ view of the evolution of ‘specialties’ in economics. Bibliographic coupling relies on cited references to determine the proximity of the citing documents: the more citing documents have cited references in common, the more they can be considered as cognitively similar. Claveau and Gingras later identify the most cited contributions for each specialty in their respective time-windows through citations counts. As we shall see, this procedure is much different from the technical definition of a historical root as identified through RPYS.

5 Agricultural Economics, American Journal of Agricultural Economics, Ecological Economics, Energy Economics, Energy Journal, Environment and Development Economics, Environmental & Resource Economics, Journal of Agricultural and Resource Economics, Journal of Agricultural Economics, Journal of Environmental Economics and Management, Land Economics, Marine Resource Economics, Resource and Energy Economics, and Resources Policy.

6 American Journal of Agricultural Economics, Ecological Economics; Environment and Development Economics, Environmental and Resources Economics, Land Economics, Land Use Policy, and Journal of Environmental Economics and Management.

7 See also the criticisms of Teixeira and Castro e Silva’s (2015) methodology on the website (accessed February 9, 2018): http://www.economics-ejournal.org/economics/discussionpapers/2015-52.

8 As a consequence, our results will show the historical roots of recent contributions to EE. It is thus only by making the assumption (which stems from our literature review) that EE is the most representative of the field that we can consider that our results can be interpreted at the field – not only the journal – level.

9 500 records at a time can be exported from WoS. The dataset was thus downloaded in 8 packages with citing references in descending chronological order, with full record and cited references.

10 We retrieved the articles without cited references by importing WoS text files in Microsoft Excel® and finding the instances where the WoS category NR (cited reference count) equals 0. Note that even though we manually detected and deleted the contributions with no cited references, CRExplorer deletes them automatically during its data import process.

11 To find duplicates, we uploaded the 8 WoS files in CRExplorer (more on this later), then we exported all the citing references in CSV format, imported the file in Microsoft Excel®, and finally used the Microsoft Excel® function of highlighting duplicates.

12 The four articles we removed were classified as ‘article’ only in WoS ‘document type’ category, so there are still 127 articles also classified as proceedings paper.

13 CRExplorer can be freely used online or downloaded at http://andreas-thor.github.io/cre/. Other tools can be used for RPYS: RPYSi/o by Comins and Leydesdorff (2016); RPYS.exe which can be found at Leydesdorff's website (http://www.leydesdorff.net/software/rpys/); Metaknowldege which is a Python 3 package with a RPYS function (McLevey & McIlroy-Young, 2017); and finally the R tool ‘bibliometrix’ developed by Aria and Cuccurullo (2017) who have very recently introduced a RPYS function.

14 Count of cited references in this section is after CRExplorer finding of duplicates but before disambiguation – more on this point later.

15 According to Andrés (2009, p. 4):

The term ‘big science’ refers to large-scale instruments and facilities, supported by funding from government and international agencies, in which research is conducted by teams or groups of scientists … ‘little science’ … was traditionally carried out by a small group of erudite scholars who then became eminent in their field of study. In comparison, ‘big science’ is characterised by large amounts of money being invested in personnel and infrastructure.

16 There is indeed a steep increase in the number of citations to cited references in our dataset, starting at the end of the fifties and ending in 2007 when the number of citations to cited references decreases steeply.

17 191 cited references were removed in the period 1215–1879 and 120,523 from the period 1951 to 2017. Another, more straightforward, way to obtain those 892 references is to define the 1880–1950 range for the cited references year of publication in CRExplorer's options of import process, excluding the cited references with no publication year. We did not follow this procedure because we wanted to provide the details of the number of cited references deleted in each step.

18 The Levenshtein distance is between 0 (perfect difference) and 1 (perfect similarity).

19 We should also mention the rare existence of mistakes in WoS extraction process and well-known issues in WoS data quality. For instance, we found – by chance events – one reference to Mauss (2000) in Poe, Levin, Tolimieri, and Norman’s (2015) article which appeared as Mauss (1924 – date of first publication) in WoS exported text file. We communicated the error to Clarivate Analytics which corrected it not in the article's bibliography but in WoS exported text file. In other words, in case of replication, there is, as we are writing this note, a reference to Mauss 1924 replaced by one to Mauss 2000. We also found 2 publications in our dataset for which the figure attached to WoS ‘NR’ tag (number of cited references) was not equal to the actual number of the cited references. In each case, the difference was of 1 missing cited reference in the WoS text file. We checked with the original articles that the 2 missing cited references were published after 1950 so this issue does not influence our results. Finally, in rare cases, there are duplicates in the cited references of a publication. CRExplorer automatically deletes the duplicates in the list of the cited references for each publication in importing the data. These two elements (incorrect figure in WoS ‘NR’ tag and duplicates in cited references for a publication) explain why there is a small difference between the sum of the figures attached to WoS NR tags and the number of cited references imported into CRExplorer. Finally, corrections made through time by Clarivate Analytics and other variations in WoS data export process influence the reproducibility of results.

20 Errors can come from authors or WoS’ data extraction process – see also footnote 19.

21 In the description of the most influential peaks, we cite as many contributions as necessary to obtain at least 50% of the total number of citations achieved for the year considered. Also, recall that in the analysis of the peaks it is impossible to consider the various editions of books, but we considered them in our calculus of global citations.

22 Accessed January 11, 2018. https://www.journals.elsevier.com/ecological-economics/.

23 That the picture we get of EE changes when considering either the five most cited contributions or the ten most cited contributions is a reminder of the classic issue faced in every bibliometric study of how to determine the optimal threshold for the number of instances to be included in any ranking. We will also encounter this caveat when studying pluralism in the next section.

24 Hicks famously characterized his intellectual evolution in those terms:

Clearly I need to change my name. Let it be understood that Value and Capital (1939) was the work of J. R. Hicks, a ‘neoclassical’ economist now deceased; while Capital and Time (1973) – and A Theory of Economic History (1969) – are the work of John Hicks, a non-neoclassical who is quite disrespectful towards his ‘uncle’. (Hicks, 1975, p. 365)

25 There have been debates on Knight's relationship to neoclassical and institutional economics with opposing conclusions (e.g. Asso & Fiorito, 2008). As only one illustration, Emmett (2009, p. 87) notes that Knight's

discussion of economic rationality … is characterized by the tension he sustains between a strong statement of the scientific validity of the neoclassical maximizing assumption and an equally strong statement of the assumption's inadequacy, both empirically and ethically, as a description of actual exchange behavior.

However, what matters for us is Knight’s (1921) book and, in this respect, Emmett (2009, p. 63) unambiguously puts Knight on the neoclassical side:

the rise of neoclassical economics during the interwar period is, in no small part, directly attributable to Knight. In fact, many consider Knight the quintessential American Twentieth-century neoclassical economist. [Risk, Uncertainty and Profit] defined neoclassical economics and ensured its acceptance.