The History of Positivism (Pt. II): Mill and Mach

(Note: This will make more sense to those who read the preceding post.)

The first dedicated non-French readers of Comte's Course of Positive Philosophy, which was discussed in the previous post, were English Humeans who rejected speculative metaphysics to the point of redefining cause as the constant conjunction of successive observations; and followers of Bentham who, like Comte, held that social progress must be made not by appeals to "natural rights" and other metaphysical ideas, but rather by applying scientific principles. JS Mill was foremost among the Humean Utilitarians and he did much to popularize Positivism, including writing a book on Comte. JS Mill had been home schooled by his father James Mill (who did much to define utility for economics) and his friend Bentham who was busy inventing (among other things) a "science of morality" based on utility construed in terms of maximizing pleasure and minimizing pain for the greatest number of individuals possible-- a form of Utililtarianism which Mill would later criticize as flawed and set about revising.

The young JS MIll was so impressed by Comte's Course On Positivism, that in 1840 he commenced a correspondence with Comte which lasted until 1847 (when the increasingly disturbed Frenchman assumed the role of secular priest in his own "Religion of Humanity" for which Mill had no patience, describing it as "spiritual despotism"). Mill's book on Comte (1865) contains 2 essays: the first a glowing endorsement of his Positivist philosophy of science, the second a harsh critique of Comte's later hierarchical "Religion of Humanity" with its cult-like structure. The early (1840-1841) letters to Comte begin with Mill calling himself a follower of Positivism, but soon it becomes clear that he was a critical follower. Mill was of a more liberal bent on social issues, but more important, he thought that Comte's omission of psychology from his classification of the sciences was ill-advised. Mill in his letters and his later book, Auguste Comte & Positivism (1865) stated that psychology was capable of becoming experimental along Humean lines no less than Sociology. If anything, Mill insisted, empirical and experimental psychology was the logical starting point for a scientific experimental method in a post-metaphysical, Humean tradition like that which Comte and he now shared. The problems discussed in the last post (i.e. introspectionism and mentalism in psychology) were to be remedied by the application of Mill's Inductive Methodology for Experimental Science, as articulated in his most important book, System of Logic. Mill's experimental psychology in no way violates the Comtean requirements of eschewing metaphysics and conforming to empirical strictures established by Hume and Comte.

In effect, Mill was carrying the positivist tenets to their logical conclusions which were the avoidance of all speculations about causation, and the introduction of a philosophy of science which complemented Comte's. More than 100 references to Comte's course are found in his System of Logic (1842). Though I can't summarize the whole book and all the "canons" of inductive logic contained between its covers, it is enough to give a general sense of the book (which went through many revisions until 1872 before Mill's death). It left a great imprint on Empiricism generally, and developed a line of thinking that was close to Mach's subsequent version of empiricism and the general positivist view of science.

Science, for Mill, can give us knowledge only within the bounds of experience. This is taken to the extreme point of redefining Matter not as a substance (which is a metaphysical posit) but as "the permanent possibilities of sensations." That is, reality (the external, material world) exists only insofar as it is the subject of experience, a position that Mach will take, and which The Vienna Circle of positivists in the 20th century will inherit in only slightly reworked fashion.. The trick, then, is to move from discrete sensations to objective and general knowledge of the external world. It is here that Mill introduces "canons of induction" in The System of Logic. For example, one such canon is called the principle of concomitance. Here A is the cause of B when in presenting A you can produce B. As with Hume we don't posit some occult force called "cause" but rather we pinpoint constant relations. As to cause, Mill defines that concept in terms of all the conditions necessary to produce an event (which we call the effect). Usually, scientists ignore most of the necessary conditions ("background conditions") to highlight one variable of importance for a hypothesis, but the notion of one phenomenon in itself serving as a cause for another phenomenon is almost entirely rejected by Mill. Thus in "The method of agreement," the experimental scientist must record all factors that are present whenever B occurs in our above example (it is not clear this is always possible). The method of difference looks to see if there is any event which is always the case whenever B occurs, and thus whether B never occurs in the absence of this factor. These and other Millsian methods are used to this day in empirical experiments of various sorts (e.g. drug trials, experimental psychology, social research, biological studies, et al.). Induction cannot be avoided in Empirical Science, and Mill did his best to move from induction to objective general truths. In his earlier editions of the book, he claimed we could move from particular sensory associations to Universals. It is not always appreciated, however, that such claims became humbler and humbler through the revisions, until in his last version in 1872, Mill wrote that: "In matters of evidence, as in all other human things, we neither require, nor can attain the absolute. We must hold even our strongest convictions with an opening left in our minds for the reception of facts which contradict them... Whatever has been found true in innumerable instances, and never found to be false after due examination in any, we are safe in acting on as universal provisionally (1872, III.xxi.4; 1973, p. 574 [emphasis added] Quoted in Duchayne; 2016). Empirical Verification is not absolute, then; a lesson that would be re-learned by later positivists in the 20th century with regard to Verificationism in empirical science, and frustrations in developing reliable inductive logics to validate scientific laws. So with Mill, the physical world is understood only in terms of experience which, it turns out cannot give us foolproof knowledge.

Mach continues in this "radical empiricist" tradition to use James' apt term. He writes in Analysis of Sensations, "We know only one source which directly reveals scientific facts-- our senses." With Mach we have reached the most direct influence on the early Vienna Circle of logical positivists centering around Morris Schlick, Carnap, Otto Neurath and others, who organized the Ernst Mach Association in the early 1920s, by which time Mach himself had died (1916). Like the Vienna Circle he so influenced, Mach railed against all metaphysical conjectures, saying of reality that it can be nothing other than that which is contained in a complete description of sense experience (c.f. Mill's definition of material reality as a "permanent possibility of the sensations"). Anything else goes beyond plausible evidence. Thus science should be redefined as an account of the facts given in sense experience. These sensations are called "elements" by Mach, and they are not to be confused with "subjective experiences" that lack objective status. Rather Mach, a physicist by training and vocation, was interested in sense data as reals not "mere appearances" as sometimes assumed mistakenly. Using the example of color, one might discuss color perception in terms of "subjective" relations such as how a color percept affects mood; but at the same time one could understand the "psychophysics" of color perception (phrased in terms of reflectancy of light, the mechanisms of the retina and so on). Like Bertrand Russell's "neutral monism," sensations in Mach's sense are not simply Humean events but psychophysical reals capable of experimental investigation and scientific description in their own right, as Mill anticipated in his System of Logic where experimental psychology was to replace intuition with investigation, and introspection with positive inductive science. There is one difference, though, between Mach and JS Mill and other empiricists. When we investigate sensory observations he finds that we perceive all objects in relations to others automatically. We don't just see a single patch of color, but we already see it against a background or upon a surface. Rather than seeing all percepts as fixed and isolated bits we relate to one another consciously, he thinks much of our empirical experience is already structured in terms of what will later be called "Gestalts" or relations like part/whole, back/front and others which he thinks have evolved physiologically to help us navigate the environment. (see Stanford Encyclopedia of Philosophy entry on Mach for thorough description of his theory of perception.) What positivists took from Mach, however, was not so much this last emphasis on relational experience, but the idea that in principle a scientific theory could be reduced to a complete description of sense experience. With the help of logical analyses of such sense data, Logical Positivists would attempt to do just that in the 20s and 30s.

However, Mach was skeptical regarding the existence of all non-observables, including atoms. If something is not given to us in the form of direct experience then it is a metaphysical posit, not a real object. What really upset him was the idea that atoms were originally seen as irreducible and isolated entities (a-tomos in Greek means "uncuttable" which we now know is false, as indeed atoms can be split, and are not "irreducible" isolates). Still, later positivists would have to come to terms with "Indirect Empiricism" such that unobservables can be counted as full-fledged empirical objects if they are indirectly or inferentially known by means of logic or else such technological instruments as electron microscopes, radio telescopes, interferometers etc. When I look at a penny and say "This is a copper coin" that fact is not part of the sensory experience taken alone. It must be joined with other statements about how we can check to see if a metal is copper. That determination is made not by simply looking at the coins, but rather by further testing of the metals. So though it is correct to say "This penny is copper" it is known indirectly through additional statements about how we identify copper. Other elements in science cannot be observed directly because they are invisible to the naked eye. These require special techologies to be detected, and are thus also known indirectly. But we will take up the transition from Empiricism based on direct observation to "Indirect Empiricism" when we discuss the Vienna Circle of logical positivists in the next post.

-----------------------------------------------------------------------------------------------------------------------------------------

References and Links to Online Resources:

-John Stuart Mill; Auguste Comte And Positivism (1865) ; Pdf here: http://library.umac.mo/ebooks/b21819853.pdf  and free Kindle download here: https://www.amazon.com/Auguste-Comte-Positivism-John-Stuart/dp/1512324248/ref=la_B000APZ4H4_1_14?s=books&ie=UTF8&qid=1483858738&sr=1-14&refinements=p_82%3AB000APZ4H4

-John Stuart Mill; System of Logic (excerpted in Ninteenth Century Philosophy: Readings in the History of Philosophy; Free Press, 1969)

-Ernst Mach; Analysis of Sensations (ibid)

-Erik C. Banks: “Machian Elements and Psychophysical Relations”in S. Mori, et al. (eds.), Proceedings of the 23rd Int'l Conference in Psychophysics: International Society of Psychophysics: 2007

-Ernst Mach entry at Stanford Encyclopedia of Philosophy: https://plato.stanford.edu/entries/ernst-mach/#Sum

-JS Mill: A System of Logic; Vol I, 8th ed. Free Kindle version available here: https://www.amazon.com/System-Logic-Ratiocinative-Inductive-Vol/dp/B003YMMOTG/ref=sr_1_32?s=books&ie=UTF8&qid=1484005293&sr=1-32&refinements=p_27%3AJohn+Stuart+Mill--  Relevant readings: Book 3 (On Induction) Ch's 8 (Methods) and 9 (Examples of methods used in science). Book has active Table of Contents. It is also available at Internet Archive as a PDF.