1“I could wish that all the things I have published on these philosophical subjects were read in the order wherein I published them [...] to take in the design and connection of them,” George Berkeley wrote to his American friends.  He had three works in mind: the Theory of Vision, the Principles, and the Three Dialogues. My suggestion about the “connection of them” is that he uses the analytical method in the Theory of Vision (as he actually says that he did) and that he presents the synthesis in the two other books. “Analysis” and “synthesis” are here taken in the old sense, according to which a non-deceiving god causes those ideas in us which we perceive “clearly and distinctly.” The main concern of the analysis was to discover these god-given epistemic atoms and the elementary relations between them. In the next step, “the Conclusions in the Analysis [are] assumed as Principles in the Synthesis” (TVV 38). If I am right, then the “Survey of the Objects of Human Knowledge,” presented in the opening sentence of the Principles, is such a conclusion based on the analysis of the Theory of Vision.
2There is a difficulty with this suggestion, however. According to a long tradition in Berkeleian scholarship, we are supposed to interpret Berkeley’s philosophy in the light of his early, unpublished manuscripts with the result that certain parts of the “Survey” appear “philosophically impossible” just as the Theory of Vision appears to be an unreliable source of Berkeley’s “true” philosophy. 
Opposing this tradition, I will argue that there is a dramatic change in Berkeley’s early development worth the name of a paradigm shift. From the perspective of the Old Paradigm, the Theory of Vision appears awkward and parts of the “Survey” appear “impossible”; but in the light of the New Paradigm, they are important, coherent parts of a surprisingly modern account of perception. If correct, then this casts new light on the role of the Theory of Vision in Berkeley’s philosophy. In this paper, however, I confine myself to the opening section of the Principles, interpreted as a condensed summary of an important theme of the analysis in the Theory of Vision.
Berkeley’s Early View on Perception
Towards a New “Science of Vision”
4I start with a general overview of Berkeley’s steps towards what I take to be his new doctrine. The first case I mention is an optical experiment. It was originally performed by Isaac Barrow to show how we perceive distance, but it failed in a way that caught Berkeley’s interest. In the experiment, an object was projected onto the subject in such a way that it looked fuzzy. Berkeley asked: How do people interpret fuzziness, when they judge about distance? In his version of the experiment, I call it “Berkeley’s Experiment,” he used two subjects, the one with normal eyesight, the other nearsighted. The normal-sighted person said “It is near,” the nearsighted “It is not near; it is far away.” I shall use the term “sensation” for what both of them saw (a fuzzy-looking object), and the term “perception” for what they saw, when the one said “I see a near-by object,” the other “I see a far-away object.” Thus in this case we have:
8In Berkeley’s Experiment, for instance, the optician observed that the rays of light did not focus on the retina, which made the physiologist predict “Confusedness of Appearance.” The psychologist could then forecast that the near-sighted person would perceive a far-away object, the normal-sighted one a close-by object.
“The Beings […], which exist without, may indeed concern a Treatise on some other Science, and may there become a proper Subject of Inquiry. But, why they should be considered as Objects of the visive Faculty in a Treatise of Optics, I do not comprehend” (TVV 19, my emphasis): “To the absolute Nature, therefore, of outward Causes or Powers, we have nothing to say” (TVV 12).
11He actually maintains that, even if his metaphysical doctrine is mistaken, then this would in no way affect his theory of vision (TVV 20).
12But how can a Black Box Theory of Mind help us to investigate mental states in another person’s mind?
Investigating Mental States in Another Person’s Mind
“We see Shame or Anger, in the Looks of a Man… for no other Reason, than barely because [certain Colours, and alterations of Countenance, which are the immediate Object of Vision] have been observ’d to accompany [Shame or Anger]: Without which Experience, we shou’d no more have taken Blushing for a Sign of Shame, than of Gladness.”
15This example illustrates the crucial pre-behaviouristic idea in Berkeley’s Black Box Theory of Mind: we know mental states only by observing a person’s voluntary or involuntary behaviour (TV 9-10).  What we see, when looking at a face turning red, is certainly no copy or image of the mental state in that person’s mind; there is no resemblance between a red face and a passion. Yet, we do perceive passions, according to Berkeley, but not by sense but by a primitive form of induction.
16Once we have noted the connection between red faces and a certain behaviour, the sensation of a face turning red begins to raise expectations in us (TV 23); we begin to take “Blushing for a Sign of Shame.” This relation between a sign and what it signifies makes Berkeley distinguish between “two sorts of Objects apprehended by the Eye: The one, primarily and immediately, the other, secondarily and by intervention of the former” (TV 50, 54). I have referred to these two kinds of “objects” as “sensations” and “perceptions.”
“Sign,” as a technical term in the Theory of Vision, is closely connected with the concepts of “coexistence” and “suggestion.” When different ideas have been “observed to go together,” or to “coexist,” we expect them to appear together even in the future.  In Berkeley’s words: “Ideas, which are observed to be connected with other Ideas, come to be considered as Signs, by means whereof Things, not actually perceived by Sense, are signified or suggested to the Imagination.” 
“Sitting in my Study I hear a Coach drive along the Streets. I look through the Casement and see it. I walk out and enter into it. Thus, common Speech wou’d incline one to think, I heard, saw, and touch’d the same Thing, viz. the Coach. It is, nevertheless, certain, the Ideas intromitted by each Sense are widely different, and distinct from each other; but having been observed constantly to go together, they are spoken of as one and the same thing”.
These observations indicate that perceiving requires memory and imagination and is a mental act, which includes both induction and expectation. It can be illustrated as follows:
Berkeley’s Constructivist Approach to Perception
20Berkeley uses the Law of Specific Sense Responses to identify what kind of objects is given. It states that,
In negative terms, the Law of Specific Sense Responses states that “the Ideas intromitted by each Sense are widely different, and distinct from each other” (TV 46). In other words:
23Berkeley regarded it the main issue of his psychology to explain, how different raw data, or pure sensibles, become interpreted as qualities of more complex units of mixed sensibles, which are “considered as one thing” and “marked by one name” (TV 14, 35, 41, 42, 52, my emphasis). “The Solution of this Problem, in its full Extent, doth comprehend the whole Theory of Vision,” he says, and adds optimistically: “Thus stating of the Matter placeth it on a new Foot, and in a different Light from all preceding Theories” (TVV 42). In (7) and (8), I have illustrated the steps from sensation to perception, and in (9) I have mentioned those pure sensibles that are “given,” according to Berkeley. But by what laws does the mind combine these raw data into more complex units?
24In Principles, 146, we read about the “Laws of Pain and Pleasure, and the Instincts or natural Inclinations, Appetites, and Passions of Animals.”  In the Theory of Vision, this law takes the form of a Preservation Principle:
This explains the direction of the process from sensation to perception. It says that,We regard the Objects that environ us, in proportion as they are adapted to benefit, or injure our own Bodies, and thereby, produce in our Minds the Sensations of Pleasure, or Pain […] Which Foresight, how necessary it is to the preservation of an Animal, every ones Experience can inform him.
27This pragmatic approach is further emphasized in the Constructivist Thesis:
“Every Combination of Ideas is consider’d as one thing by the Mind, and in token thereof, is mark’d by one Name. Now, this Naming and Combining together of Ideas is perfectly Arbitrary, and done by the Mind in such sort, as Experience shews it to be most convenient. Without which, our Ideas had never been collected into such sundry, distinct Combinations as they now are”.
29This is an important theme in the Theory of Vision. Without “the slow Steps of Experience,” pure sensibles “were not connected in our Minds.” And when we have learnt to use pure sensibles as signs for creating useful “Combination of Ideas,” then Berkeley asks us to “observe that Signs are variable and of Human Institution” (TV 144). As we read in the quotation above, this “Combining together of Ideas is perfectly Arbitrary” in the sense that it is not necessary, which means not god-given (TV 109).  One important consequence is what I call the Heterogeneity Principle:
(hp) Ifx is a mixed sensible, i.e. a unity of heterogeneous elements, then x is not given.
31Several interesting consequences follow from the thesis that the mind combines ideas into units “as best suits it’s own Ends and Purposes.” If we “never” regard a “Combination of Ideas” as “one thing,” unless we regard it “most convenient” in the present situa-tion, then one consequence is the Relativity Principle, which says that:
Psychologists and zoologists will not be surprised at this standpoint, but philosophers will probably ask whether the faculty of perceiving is supposed to serve pragmatic purposes rather than to get at the truth? This question illustrates a fundamental difference between Berkeley’s early metaphysics and his “Science of Vision.” To the Early Berkeley, there was an immediate connection between “reality” and “truth,” and the connecting link was god-given perceptions. So, at that early stage, our perceptions did provide him with indubitable “truths.” But his “Science of Vision” is not about ontological truths; it is about the selective process in which animals tend to pay attention to some things, neglecting others, in the light of certain, probably inborn, pragmatic criteria (TV 147, cp. TV 85, 87, P 146).(rp) If P and Q live under different conditions, with radically different background-knowledge, and P perceives x as “one thing” in S, then Q could very well be quite unable to perceive x as “one thing” in S.
The Paradigm Shift
32The Early Berkeley, on the one hand, spoke about “objects” as (1) “real” things, which (2) we are acquainted with by virtue of god-given perceptions; (3) and these perceptions represent “real” objects by resemblance.
33In the Theory of Vision, on the other hand (1) he is not talking about “objects” in the everyday sense of “real” things; ontological issues are irrelevant to his “Science of Vision”; (2) only such “objects” as pure sensibles are “given,” whereas perceptions are end products of mental processes; (3) and these perceptions do not represent what they signify by likeness.
34When the term “object” is used as a technical term in the one of these two contexts, its meaning is incommensurable with its use in the other context: a thesis formulated in the one cannot be translated into the language of the other. Unfortunately, Berkeley does not always use “object” in a technical sense, which creates a serious terminological difficulty. When he presents what he is about to explain, then he speaks about “objects” in a non-technical everyday sense; but when he starts explaining, or developing the issues, then he speaks about “objects” in technical senses that are incompatible with everyday usage.  He admits, he had been “betrayed thereby to say things, strictly speaking, neither true nor consistent.” On the other hand, it is quite easy to follow his arguments, if we observe his method to proceed step by step, or “by degrees,” as he puts it; first to formulate the problems in non-technical terms, or outside the theory, “gradually correcting our Judgment, and reducing it to a Philosophical Exactness” towards the end, when all entities have been properly analysed, defined, and brought within the theory (TVV 35).
If a paradigm shift of the kind I have outlined can be established, then it can explain, at least, the classical difficulty in understanding the opening section of the Principles.
The “Survey of the Objects of Human Knowledge”
35Berkeley opens the Principles (Part I) as follows:
“It is evident to any one who takes a Survey of the Objects of Human Knowle[d]ge, that they are either Ideas  actually imprinted on the Senses, or else such as are  perceiv’d by attending to the Passions and Operations of the Mind, or lastly Ideas  formed by help of Memory and Imagination; either compounding, dividing, or barely representing those originally perceiv’d in the aforesaid ways (P 1, numbers in square brackets are mine).”
37In the light of the Old Paradigm, the second kind of ideas appears impossible; we cannot have an image-picture of a passion or of a spirit operating about its ideas.  With the exception of a few scho-lars, who maintain that Berkeley had Lockean “ideas of reflection” in mind, an overwhelming majority of the experts argue that Berkeley did not mean what he says here. 
38But Berkeley does not invite us to look into the Black Box to see how the soul is operating about its ideas, or to observe what a passion in another person’s mind would look like. I agree with the old authorities: that would indeed be awkward.  In the Theory of Vision, Berkeley uses ideas of passions to show that “ideas” are no pictures but operative elements in perceptual processes (TV 9, 10, 23, 41, 65, 94), and the main problem of this work was to show how the mind operates about ideas in its “Combining together of Ideas” into more complex units.
39In section one of the Principles, Berkeley refers to such combinations of ideas by the term “sensible thing.” His example is an apple : “a certain Colour, Taste, Smell, Figure, and Consistence having been observ’d to go together, are accounted one distinct Thing, signified by the name Apple.” Such a full-sense compound of coexisting heterogeneous elements is not given, according to Berkeley’s psychology; and it is not perceived by sense but by the imagination. The new question, when he returns to ontological issues in the Principles, is: Can we project our perceptions onto “reality”?
To sum up. If I am right about Berkeley’s development, then he took it for granted at first that our perceptions were god-given pictures of reality:
- “Reality” ? Perceptions.
In the next place, he began to distinguish between what I called “sensations” and “perceptions.” The result was a pioneering work on perceptual processes:
- Sensations ? [ – – ? – –] ? Perceptions.
When he returned to ontological issues in the Principles, he started from the results of his psychological investigations and formulated a new problem than can be illustrated as follows:
- Perceptions ? [ – – ? – –]. How do our perceptions relate to the “reality”? But that is a question that falls outside the scope of the present paper. 
A. A. Luce and T. E. Jessop (eds), The Works of George Berkeley Bishop of Cloyne, 9 vols (London: Thomas Nelson & Sons Ltd., 1948-1957). Hereafter referred to as the Works. The letter is quoted from vol. 2, p. 294.
See the editorial comments in Works, vol. 1, p. 149-150, 156, and vol. 2, p. 8-9, 41 n.
See the early part of Notebook A, “Of Infinites” and the Manuscript Introduction (MI).
I use clumsy arrows, “?,” non-technically to illustrate what happens over a period of time.
Quoted from the editorial comments in Works, vol. 2, p. 8-9, 41 n.
William Molyneux, Dioptrica Nova (London: Benj. Tooke, 1692), p. 106. Italics original.
In this quotation from TV 65, “certain” in “certain Colours” is my gloss.
When Berkeley, later on, became convinced that there is a spirit in the box, he continuously said: “Such is the Nature of Spirit, or that which Acts, that it cannot be of it self Perceived, but only by the Effects which it produceth” (Principles, 27).
TV 25, 55, 66, 72, 103, 110, 145.
TVV 39. TV 25, 26, 28, 36, 37, 47, 50, 51, 53, 57.
In the Principles, Berkeley refers to these “never enough admir’d Laws” in an argument from design for the existence of a wise and good creator.
For arbitrariness in this sense, see also TV 143: “It is indeed Arbitrary that, in general, Letters of any Language represent Sounds at all; but when that is once agreed, it is not Arbitrary what Combination of Letters shall represent this or that particular Sound.”
We read, for example, in the opening section that his design is “to shew the manner, wherein we perceive by Sight the Distance… of Objects” and, accordingly, he mentions “An Object placed at a certain Distance from the Eye” (21); but then we read that “things placed at a Distance are not, strictly speaking, the Object of Sight” (46), that a “proper and immediate Object of Sight” is “not without the Mind” (43), and that people are prejudiced “in thinking what they see to be at a distance from them” (43).
Some scholars argued that it is a “grammatical mistake” to take “such as” in Berkeley’s “survey” to refer to “ideas”; it does refer to “objects” (G. A. Johnston, The Development of Berkeley’s Philosophy (London: Macmillan, 1923, p. 144). But as Alfred Klemmt argued in Klemmt, 1957, p. 122, and
E. J. Furlong, in “An Ambiguity in Berkeley’s Principles,” Philosophical Quarterly, 14 (1964), 334-344: “such as” refers to “ideas.” Furlong observes that the sentence ends in a reference to “ideas… perceived in the aforesaid ways.” As “ways” is in the plural, it “must refer to the two previous classes specified, namely (i) ‘ideas actually imprinted on the senses’, and (ii) ‘such as are perceived by attending to the passions and operations of the mind’.” But, Furlong adds, “philosophically” he must have meant “objects,” not “ideas” (335).Online
Those, who suggest that Berkeley adopted John Locke’s ideas of reflection are A. C. Fraser, Editor’s note, in Works, vol. 1 (Oxford, 1901), p. 257; Alfred Klemmt, “Anmerkungen,” in George Berkeley. Eine Abhandlung über die Prinzipien der menschlichen Erkenntnis (Hamburg: Felix Meiner, 1957), 122; Daniel E. Flage, “Berkeley’s Ideas of Reflection,” in Berkeley Newsletter, 17 (2006), 7-13. For references to authors who argue that Berkeley did not mean what he said, see the paper by Daniel Flage.
From the assumption that “Berkeley always means by ‘idea’ a sensory object,” and “sensory objects are essentially concrete, particular, picturable,” Works (2:8-9, 41n), the second kind of ideas in particular caused serious difficulties. One of the editors declared, for instance: “It is clear to me that the objects we perceive ‘by attending to the passions and operations of the mind’ are ourselves, other finite spirits, and God.” No one accepted the absurdity that we could look at a soul and its passions and observe how this soul operated about its ideas. The dogmatic conclusion was: “Ideas perceived by attending to the operations of the mind do not exist for Berkeley.” See A. A. Luce, Berkeley’s Immaterialism (London: Thomas Nelson & Sons, 1945), p. 39f.
I wish to express my gratitude to Geneviève Brykman and to the Hultengrens Fond, Lund University, for a grant that made it possible to write this paper.