Wholism, reductionism (Francois, 2004)

Proponents of #SystemsThinking often espouse holism to counter over-emphasis on reductionism. Reading some definitions from an encyclopedia positions one in the context of the other (François 2004).

–begin paste —

HOLISM 1) – 3)
“A descriptive and investigative strategy which seeks to find the smallest number of explanatory principles by paying careful attention to the emergent properties of the whole, as opposed to the behavior of the isolated parts, as chosen by the observer in a reductionist strategy” (T.F.H. ALLEN & T.B. STARR, 1982, p.270).

The term and the concept were introduced in 1926 by the South African general and statesman Jan SMUTS. The term was derived from the Greek: “holos” = whole.

SMUTS wrote: “The idea of wholes and wholeness should… not be confined to the biological domain: it covers both inorganic substances and the highest manifestations of the human spirit. Taking a plant or an animal as a
type of a whole, we notice the fundamental holistic character as a unity of parts which is so close and intense as to be more than the sum of its parts; which not only gives a particular conformation or structure to the parts, but so relates and determines them in their synthesis that their functions are altered; the synthesis affects and determines the parts, so that they function towards the whole; and the whole and the parts therefore reciprocally influence and determine each other, and appear to merge more or less their individual characters: the whole is in the parts and the parts are in the whole, and this synthesis of whole and parts is reflected in the holistic character of the functions of the parts as well as of the whole” (1926-1973, p.86).

M. BUNGE describes as follows the characteristic theses of holism, “the ontological view that stresses the integrity of systems at the expense of their components and the mutual actions among them”:

“1. The whole precedes its parts.
“2. The whole acts on its parts.
“3. The whole is more than the sum of its parts.
“4. Wholes emerge under the action of agents that transcend both the actions among the components and the environmental influences.
“5. Totalities cannot be explained by analysis: they are irrational.
“6. The whole is better than any of its parts” (1979, p.39-40).

BUNGE sharply criticizes these thesis, reproduced however by him in a somewhat caricatural form (see hereafter)

According to ALLEN and STARR, both holism and reductionism seek to explain emergent behavior by invoking a lower level of organization (p.270).

Thus, both strategies admit the existence of hierarchies in systems.

J .A. GOGUEN and F.J. VARELA however observe: “Most discussions place holism/reductionism in polar opposition. This seems to stem from the historical split between empirical sciences viewed as mainly reductionist or analytic, and the (European) schools of philosophy and social sciences that grope toward a dynamics of totalities” (1979, p.40).

However: “It seems that both these directions of analysis always coexist, either implicitly or explicitly, because these descriptive levels are mutually interdependent for the observer. We cannot conceive of components if there is no system from which they are abstracted, and there cannot be a whole unless there are constitutive elements” (p.41).

These authors give the excellent example of harmony and melody, which are at”… a level of organization above that of the notes them selves” (Ibid).

Finally: “Reductionism implies attention to a lower level, while holism implies attention to a higher level. These are intertwined in any satisfactory description; and each entails some loss relative to our cognitive preferences, as well as some gain” (p.42).

M. BUNGE holds a dim view on holism, which he carefully distinguishes from systemics, as holism “recognizes the existence of systems with specific characters (emergent properties), but treats them as totalities or black boxes“. According to him holism “… refuses to analyse them and to explain the formation and the collapse of wholes in function of their components and the interactions between them” (1995, p.16). He also indicts holism as “responsible for the backwardness of the non-physical sciences. It has contributed precious little to serious systemics, precisely because: (a) it has not engaged in the study of the links that hold any system together, and (b) rather than constructing conceptual systems (theories) to account for concrete systems, it has spent itself in attacking analytical or atomistic approach and praising totality as such. Whatever truth there is in holism – namely that there are totalities, that they have properties of their own, and they should be treated as wholes – is contained in systemism, or the philosophy underpinning systemics” (p.410).

The most equilibrated view has been offered by G. KLIR who considers that present systems thinking: “… represents a synthesis of the reductionistic thesis and the holistic antithesis” (1993, p.36).

The French philosopher B. PASCAL anticipated (1670!) this view: “I consider impossible to obtain knowledge of the parts without knowing the whole, nor to know the whole without particular knowledge of the parts” (Quoted by R. VALLEE, 1995, p.11).

The concept of Gestalt, as refering to the perception of wholes, also is another conspicuous root of holism.

Some view holism itself as a kind of reductionism. K. BAUSCH for ex. defines it as “A reductionist descriptive and investigative strategy for generating explanatory principles of whole systems”(Glossary, Pers.comm., 2002)

Indeed: “Attention is focused on the emergent properties of the whole rather than on the behavior of the isolates parts”(lbid).

Of course, holistic models should be paired with classical reductionist ones as both aspects are complementary and necessary for comprehensive explanations.

As a very simple illustration, while H20 has expecific properties as a whole, its constitution can be understood only by knowing the chemical and physical characteristics of Η and 0 that allow them to combine.

— end paste —

With holism on one side, reductionism is on the other.

— begin paste —

1. “A descriptive and investigative strategy which gives account of phenomena in terms of a series of isolated parts, coupled together by direct causal linkages” (T.F.H. ALLEN & T.B. STARR,.1, p.276).

2a. A principle “according (to which) all scientific concepts are reducible to a set of ultimately irreducible concepts” (R.L. ACKOFF, 1974, p.53).

2b. “The belief that everything in the world and every experience of it can be reduced, decomposed, or dissembled to ultimatly simple elements, indivisible parts” (R.L. ACKOFF, 1991, p.325).

3. The claim “that properties of a whole are explicable in terms of properties of the constituent elements” (G. KLIR, 1991, p.24).

4. “… the task to find the simplest, most economical and (usually) most elegant explanation that will cover the known data” (G. BATESON, 1979, p.230).

T.F.H. ALLEN and T.B. STARR add: “Ambiguity in relationship between parts is met with further subdivision until the ambiguity disappears” (Ibid).

As a result “Reductionism is thus the strongest possible way of ordering the list of the various sciences” (I.I. MITROFF and H.A. LINSTONE, 1993, p. 165).

Commenting his definition, ACKOFF states that these “ultimate concepts”, according to some “were provided by direct observation” and to others “thought of as undefined concepts of a formal system”. And, “Whatever
their source, these concepts were identified as “physical thing predicates”; that is, physical properties of things”.

BATESON adds to his definition the following caveat: “Beyond this, reductionism becomes a vice if it is accompanied by an overly strong insistence that the simplest explanation is the only explanation. The data may have to be understood within some larger gestalt” (Ibid).

For a good example, see: Selection (Multi-level)

Reductionism is a paradigm. R. ROSEN expresses this as follows: “The belief that any natural system can be so decomposed (note: i.e. extensively, and possibly limitlessly), and that the laws governing the motions of these
particles can be determined, is the essence of reductionism” (1979, p.174). The “belief” is something like the hardening of an assumption.

The conceptual origins of reductionism are to be found in DESCARTES and NEWTON, if we are not going back in time to DEMOCRITUS, EPICURUS and LUCRETIUS.

The basic (and largely unconscious) tenets of reductionism seem to be the following:

  • what should be researched are the properties of objects, which are “real” and perfectly knowledgeable
  • the observer or experimenter is “transparent”, i.e., without influence on the perceived objects
  • the “et ceteris paribus” postulate can be applied without restrictions, which allows for linear causal explanations
  • supposedly, no significant aspects of phenomena are thus left out, (forgetting for example interrelations between elements or parts, or between levels)
  • larger contexts do not influence upon phenomena

Systems concepts are however not opposed to the reductionist approach, as they admit levels of description. This point was made as follows by P. WEISS (quoted by G. KLIR, 1991, p.26), who describes a neutral stand: “The reductionist likes to move from the top down, gaining precision of information about fragments as he descends, but losing information content about the larger orders he leaves behind: The other proceeds in the opposite direction, from below, trying to retrieve the lost information content by reconstruction, but recognizes that that information is not forthcoming unless he already had it on record in the first place. The difference between the two processes, (is) determined partly by personal
predilection, but largely also by historical tradition”.

This does not makes clear that the main concern of the systemist are the interactions within the same level as well as between different levels of complexity, in the whole.

Moreover, as stated by I. PRIGOGINE (quoted by F. David PEAT, 1987, p.64), “… there is no “fundamental level” in nature but rather each level involves its unique description and is conditioned by the levels around it”.

In some sense, reductionism is thus “in scribed” within the systemic approach as: “Nature… requires pluralistic descriptions and … this pluralism must contain both causal and synchronistic aspects” (Ibid).

In an overview of this subject, J.A. GOGUEN and F.J. VARELA conclude: “Reductionism implies attention to a lower level, while holism implies attention to a higher level. They are intertwined in any satisfactory description; and each entails some loss relative to our cognitive preferences, as well as some gain” (1979, p.42).

Reductionism may also be understood as the use of OCKAM’s razor, as stated by P. CHECKLAND, who writes: “And we use reductionism in another sense in explanation, explaining the results using the smallest number possible number of concepts, only importing more elaborate concepts when defeated in this” (1976, p.128).

Still another meaning frequently given to reductionism is the stand according to which social and biological sciences can ultimately be reduced to explanations based on physical sciences. While anything social or biological is by necessity based on physical properties, the reductionist stand ignores the existence of succesive levels of complexity grounded on the emergence of dissipative structures and synergies.

On this topic D. CAMPBELL accepts:”… the limited emergentist principle that laws of biology, psychology and sociology exist which are not described by the laws of physics and inorganic chemistry. These emergent laws are compatible with the laws of physics and chemistry but not derivable from them” (1975, p.1104).

At long last, the fundamental difficulty becomes obvious: so-called laws must be non-contradictory between them to be acceptable on all levels. The best example has been the elimination of the vitalist stand in biology when the appearent impossibility of life in terms of classical thermodynamics was taken care of succesively by such models as open system, homeostasis and finally dissipative structuration. It should be noted that none of these models are “laws”, nor their applications limited to any particular scientific discipline.

P.M. ALLEN et al. observe that “the very success of reductionist science has… provided man with the power to radically change his environment. However, this science has afforded almost no knowledge of the probable effects of such actions in the complex systems encountered in the domains of biology, ecology and socio-economics. Policy today must be formulated in a world of ever increasing interaction and complexity” (1984, p.2).

— end paste —

The numbers beside the encyclopedia entry mean …

  • The following special markers have been used, in order to enhance the usefulness of the encyclopedia:
  • 1) meaning “systemic on a wide range”, or “general information”
  • 2) meaning “general abstract or mathematical model”, or “methodology”
  • 3) meaning “epistemologica! or ontological aspects”, or “semantics”
  • 4) meaning “practical in human sciences”
  • 5) meaning “more specific or disciplinarian”

In this paper-first encyclopedia, the bolded text is link to other entries.


François, Charles, ed. 2004. International Encyclopedia of Systems and Cybernetics | 2nd ed. De Gruyter Saur. https://doi.org/10.1515/9783110968019.

International Encyclopedia of Systems and Cybernetics

#holism, #reductionism