This book is so rich that it can only be summarized in point form, with brief comments. Square brackets [ ] indicate my addition, opinion, or comparison.

The chapter “As Every Schoolboy Knows” [or rather SHOULD know) contains some general principles of the epistemology of science.

1. Science never PROVES anything. It PROBES. i.e. Induction can fail, as shown in the example of a number series.
2. The map is not the territory, and the name is not the thing named. (Alfred Korzybski, General Semantics.) The “Ding an sich” (Popper’s World I) is inaccessible. But we must map, name, and classify [i.e. Symbolize, as Terrence Deacon says in “The Symbolic Species”. Human minds operate almost entirely in the sYmbolic system, related to language.
3. All experience is subjective, i.e. Within an individual mind. Experience is based on neuron firings in brain nets. [But neuron nets are something we derive from experience in scientific observations. Isn’t that a circular argument?]
4. Perception processes (e.g. Visual) are unconscious, though very complex. Examples:(a) Visual perspective – clues come from size and brightness; which is dominant if in conflict? (b) Parallax experiments of Adalbert Ames. (c) The trapezoidal room. (d) The anti-aircraft gunner.
5. How to divide a whole into parts is arbitrary. Example: the hexagon-triangle.
6. Divergent sequences are unpredictable. Examples: (a) Throwing a stone to break a pane of glass. (b) Where is the chain’s weakest link where it breaks? (c) Superheating water; a “seed” is needed to initiate change of phase. Same in undercooling, precipitation/crystalization, etc. (d) The great man theory of history and the theory of “social forces” (cf. Tolstoy’s “War and Peace” about Napoleon). The social forces “superheat” the system, but it does not undergo phase change until a great man precipitates it. [Did Gorbachev end the Cold War?] [This should be compared to the butterfly effect in chaos theory and to Gould’s statement that the tape of evolution, if rewound, would not repeat exactly. ]
7. Convergent sequences are predictable. [To put it in the language of prigogine in “The End of certainty”, ensembles can be predictable even if their constituent particles move randomly.] Similarly, societies can be predictable even though individuals are not. The bridge between the unpredictable parts and the predictable wholes are the laws of probability and stochastic processes. But sometimes wholes can be divergent (unpredictable, chaotic) too.
8. Nothing will come of nothing. (A quotation from “King Lear”.) This applies to the laws of conservation of energy and matter (actually their common essence), and to the principle of “omne vivum ex vivo” in biology [except for the very first beginning]. It follows that epigenesis (embryology) is subject to this principle, but evolution (tautology) is not. [“Tautology” because Darwinian evolution can be stated as a tautology: “the fit survive”, “the survivors were fit”.]
9. Number is different from quantity. Numbers come from counting, quantities are measured. Numbers can be absolutely accurate, quantities are subject to experimental error. [Accounting and book-keeping is different from physics and chemistry.] Number systems are digital, quantitative systems are analogue. [But this applies only to integers. Fractions and especially irrational numbers are not discrete, but fill the number line continuously, like a measurement.]
10. Quantity does not determine pattern. But pairs of quantities can, through ratios or differences. Example of an island with two mountains on it; if the water level rises, it can become two islands. [An example of “quantity turning into quality”.]
11. There are no monotone values in biology. “More” is not necessarily better; there is an optimum, with either a deficiency or an excess being harmful. [The Goldilocks effect.] Money seems to be monotone, but actually an excess is toxic. [This may be relevant to the existence of intransitivity.]
12. Sometimes small is beautiful. (Modified quotation from E.F. Schumacher.) An elephant and a shrew are adapted to their size; so is a dog and a flea. The sad case of a polyploid horse who could not stand up or eat enough. This comes from the fact that length grows linearly, area as the square, volume as the cube.
13. Logic is a poor model of cause and effect, because there is no time dimension in logic, while time is important in causality. The electric buzzer can operate physically as the switch is alternately closed and opened, but the Cretan liar paradox in logic, which seems to have the same structure as the buzzer, pays no attention to time.
14. Causality does not work backwards [except perhaps in quantum physics?], but logic can, again because of the time factor. So purpose and teleology are not permitted in science. [Even for conscious planning beings?] But cycles are common – thus “effect” can contribute to “cause” via feedback, positive (runaway) or negative (control).
15. Language, because of its subject-predicate-qualifier-object syntax structure, asserts that “things” have “qualities”. But in reality, qualities are relative to those of other things, never absolute. Only differences count.
16. Stability and change. For an acrobat, or an organism maintaining homeostasis, stability is the result of self-balancing change. This is the dialectical synthesis of Zeno and Heraclitus.
17. We perceive only differences. Examples are: binocular vision, the discovery of Pluto, synaptic summation in firing neurons, Macbeth’s hallucinated dagger, algebra and geometry confirming each other, the complementarity of two sexes, beats in combining sounds of two different frequencies, moire patterns, how to define “left” and “right”, description and explanation. [These examples seem to illustrate mainly complementarity, not inability to perceive single items. ]

The Criteria of Mental Process.

1. A mind is made up of parts not themselves mental. It is immanent in the organization of the parts, and emergent from this organization. Since an electron is thought to be an indivisible particle, it cannot include a mental process. [Bohm in his deterministic theory of quantum phenomena endowed an electron with a mind – receiving information about the second slit from its antenna-like wave; but he had to assume that the electron is a complex composite, not an elemntary particle.]
2. Interaction is always triggered by a difference. Our eyeballs move in vision; we see outlines, therefore we draw them; we are unaware of slow changes, e.g. In climate.
3. There has to be collateral energy entering the process, although.its entry is triggered by the mind acting as a relay, switch, gate — like turning on a water tap. [Cf. My essay on The Three Essences, which proposes a double relay: information moves energy which moves matter.]
4. Cycles are operating, characterized by positive and negative feedbacks. This is compared to the governor of a steam engine.
5. Coding the observed difference. This is the mapping-naming step, already mentioned above, which I include as symbolization. [But all minds are not sYmbolic, according to Deacon, only human minds. Others operate by iconic and indexical relationships.]
6. There is a hierarchy of logical types (classes of classes of classes…according to Russell and Whitehead). [But this is part of the sYmbolic mind system, which presumably animals do not have. Therefore I have trouble understanding his next point.] Examples of cross-species communication, e.g. Dog-ape and dolphin-human. Out of the hierarchy of logical types arises context, meaning, and understanding. It also functions in human and animal play, [which is like experimentation in acquiring meaning].

The two great stochastic processes.

1. These two processes.are genetic change and somatic learning. Lamarck was wrong in thinking that acquired characteristics are directly inherited. This would lead to too much instability (perhaps a total collapse) of the genetic system, geared to preserving structure. The weissmannn barrier between the genetic and the somatic systems in organisms prevents this. Somatic learning is transferred to the genetic level only through the operation of natural selection: those possessing a favourable trait (for a particular environment, perhaps a recently changed one) leave more descendants for the next generation. Some kind of selection process must always intervene between the conservative (stability) system and the change (flexibility) system. There is an optimum balance between them, on which selection operates.
2. Epigenesis and evolution are another example of two stochastic processes – the first frequently occurring, fast, and pre-programmed, the second composed of rare and slow steps subject to contingency. In embryonic development, the entry-point of the sperm into the egg marks the median plane of the bilateral symmetry in a frog. [Cf. My essay The Master switch.]
3. Homology and recapitulation. This led Ernst Haeckel to formulate his statement that “ontogenesis recapitulates phylogenesis”. Some homologous organs in different species represent a distortion of the underlying grid.
4. Adaptation can be compared to addiction. Both are processes of accommodation to a new ingredient in- the environment (a climate or habitat change or a new drug). But adaptation is functional and addiction is dysfunctional. [Perhaps adaptation is initially dysfunctional too, but is necessary, because the organism cannot change the environment, but we can get off the drug.]
5. Structure (form, pattern) and process (change in the former) are another pair of stochastic processes. They form a zig-zag progression in many cases when ascending to higher levels. [Cf. My essay Rise and Run.]