Our second official reading is the paper credited with introducing the term “gestalt” is the sense that would become influential in psychology, Christian von Ehrenfels (1890) “On ‘Gestalt Qualities,'” conveniently available online in English translation here.
(For those keen to read in German, the original “Über ‘Gestaltqualitäten'” appeared in Vierteljahrsschrift für wissenschaftliche Philosophie 14: 249–292.)
As discussed in our Smith reading, Ehrenfels paper appears to be a historical bridge between several different strands of thought. It builds on the ideas in Mach’s (1886) Contributions to the Analysis of the Sensations, drawing analogies between spatial and auditory perception. Looking forward, while himself Austrian and influential on the Graz school, Ehrenfels’ view appears to share key similarities with that of the Berlin school as well, which it influenced not least through Wertheimer’s early training in Prague. In particular, Ehrenfels view does not (as described by Smith) present gestalt qualities as derived by a process of production, but rather as directly perceived additional qualities.
The third reading will be posted in two weeks, but please feel free to continue to comment on earlier readings as you get around to them. Our plan is to look next at some work from the phenomenological tradition.
I found Ehrenfels to be an interesting combination of refreshing and dusty.
Some aspects of his discussion seemed quite strikingly off from a contemporary perspective, for instance his remarks on the impossibility of similarity relations obtaining between elementary sensations. I take it that the operationalization of such similarity relations in the methods of psychophysics is enough to demonstrate their legitimacy. Rather the point seems to me (and this does seem like an insight) that the type of similarity relation that obtains between elementary sensations and that between gestalten differ qualitatively — I would take the latter to be “structural” for instance, but not the former.
A pervasive feature that was refreshing, however, was Ehrenfels’ willingness to elaborate and defend foundational claims, about whether or not there can be gestalten at all, about the evidence for them in pitch perception, about the extent to which the notion of a gestalt quality over and above the sum of elementary qualities may result in a contradiction, etc.
Some questions that arose for me:
• There were places in the translation where I wondered how much connotations were affected by the choice of wording in a way that may distort Ehrenfels intent from a contemporary perspective. To pick just one example, in the discussion on p.85 of collecting sensations under a “single consciousness” or “n distinct consciousnesses.” The word for “consciousness” here is Bewußtsein, which can mean “consciousness” or “awareness” or even “sensibility” — what I can’t get a handle on (as a non-German speaker) is whether the connotations of “consciousness,” esp. those it bears in contemporary analytic philosophy of mind, are the right ones to grasp Ehrenfels’ implications here, or whether something much more deflated is meant.
Another example is the discussion of “relations” (§8), which translates “Die Relation,” yet still contains puzzling aspects that seem to go beyond the intuitive connotations of the ordinary term “relation” in English. Clearly Ehrenfels is building here on discussions of Lotze and Meinong in which “relation” has become a technical term in a way I do not fully grasp.
• A major worry about Ehrenfels’ view for me was the way in which, by the end, almost every mental construct was being classified as in some sense a gestalt (§9) — this move seems to me to emasculate the notion of gestalt so as to be almost meaningless. In particular, Ehrenfels seems to have lost sight in the discussion of “higher-order gestalten” of the key idea that gestalt are sensations of some sort.
• Ehrefels’ discussion of abstraction and even of an “Abstraktionsvermögen” seems closely in line with uses of the concept I’ve looked at in Herbart and Joh. Müller — I guess the root here is Kant, or maybe even earlier, but I would be curious where Ehrenfels stands on debates about the mechanism of abstraction and the sense to which it can be employed to defeat skeptical worries.
• Finally, I really appreciated Ehrenfels’ frequent reference to examples from art, and treatment of perceptual and aesthetic issues as closely allied. Perhaps this was the inspiration for Smith’s discussion of this relation in the previous reading; however in Ehrenfels case I interpreted the use quite differently. For Smith, I was worried that he hoped a detailed gestalt theory would explain aesthetic value; in Ehrenfels, however, I felt aesthetic sensation was being treated as a source of evidence on gestalten in perception commensurate with more mundane perceptual examples.
Why were you worried that Smith hoped a detailed Gestalt theory would explain aesthetic value (what would be worrying about that hope)?
I also found Ehrenfels’ characterisation of higher order Gestalt qualities striking. According to him, Gestalt qualities of this kind can be originated in virtue of the combination and comparison between ‘more basic’ Gestalt qualities. As Smith, in the previous text, noted, for Ehrenfels, Gestalt qualities can be combined without limit, thus bringing about qualities of higher order.
Per Ehrenfels, some conceptual complexes are, in a sense, evidence of the limitless creation of Gestalt qualities; in particular he refers to the concepts which combine “presentational contents of physical and psychical occurrences”. Ehrenfels exemplifies the former with cases such as “kindness, service, rivalry, marriage”, to name a few. (For me, it is not entirely clear how he selected these examples. It could be argued that almost any concept accomplishes such characterisation in an uninteresting way.) I believe he takes this kind of concepts to be evidence of the existence of higher order qualities because, later, he wonders whether we would have such concepts if we lacked “any embracing, unifying bond -a Gestalt quality of a higher order”. In other words, there are conceptual complexes that are preceded, and made possible, by Gestalt qualities of higher order.
The origination of higher order Gestalt qualities results from different intellectual powers, for instance, abstraction or imagination. Nonetheless, he refuses to admit that these processes of origination imply that Gestalt qualities are created intellectually. It is only because they were already “present” in consciousness during sensation that they can be brought about afterwards by other mental activities. He claims: “wherever a complex which can serve as the foundation for a Gestalt quality is present in consciousness, this quality is itself eo ipso and without any contribution on our part also given in consciousness”. Later on, he claims that these higher order qualities are already there but we are unable to “direct our attention”. Higher order qualities are somehow dormant or latent until the exercise of imagination, for instance, awakens them.
At the beginning, I thought Ehrenfels wanted to argue that sensation is conceptual; but I think he rather wants to claim that many of our concepts are already present in sensation.
The starting point of Ehrenfels is a phenomenological fact noticed and described by Ernst Mach, following which, in our experience, we have direct sense-impression [empfinden] of some complex phenomena such as melodies or spatial configurations. In this article, his original question concerns the very nature of these configurations. He proposes the following alternative: Are they “mere sum of elements or something novel in relation to this sum, something that certainly goes in hand with, but is distinguishable from, the sum of elements?” (p. 83). For my biggest happiness, this ontological question is quickly (2 pages later) refined from a more psychological perspective: “we can ask whether the spatial figure be more than the sum of the individual local determinations, whether the consciousness which apprehends the figure in question brings to presentation something more that the n individuals taken together.” (p. 85).
Being trained in the School of Brentano, Ehrenfels does know that the answer depends both on the type of the consciousness involved (perception, imagination …) and the type of figural phenomenon at stake (non-temporal, temporal, visual, auditory…). This problem requires a circumstantiated answer and a systematic investigation. As he puts it clearly: “It is now necessary to provide a survey of the manifold of Gestalt qualities and of their significance in our psychic life.” (p. 93)
But quite interestingly, just before investigating this alternative between summative vs. not summative conceptions of auditory and space shapes – Ehrenfels ends §2 by this quite surprising methodological comment with respect to a brentanian act-psychology according to which inner perception is supposed to remain the core-method to perform psychological investigations:
“The first court of appeal to which we are disposed to turn in the solution of problems of this nature – the evidence relating to the phenomena in question that is provided, obliquely, by inner perception – can here be called in aid only with difficulty. For anyone who, in dealing with such relatively subtle distinctions, is able to glean conviction from this source will normally find himself incapable of transmitting this conviction to those who are other opinion. And he will come up against the fact that many will believe themselves constrained to reject one or other of the alternatives from the very start as absurd.”
Ehrenfels explicitly recognizes the inadequacy of inner perception to investigate sensory contents and introduces a new methodological procedure as we shall deal with in one moment.
One must recall that during the first step of the Gestalt dispute (from 1890 to the beginnings of 1910s), nobody would challenge the basic assumption according to which a whole is not reducible to the mere sum of its parts. Ehrenfels is no exception. Indeed, in accordance with his definition of Gestalt quality which is introduced at the end of §4 (p. 93), Ehrenfels shows in the same paragraph why presentations of local elements are necessary but not sufficient conditions for Gestalt quality. The existence of such irreducible Gestalt quality is showed by an empirical demonstration that will later be famous under the name of “transposability”. It relies on a basic phenomenological fact that can be observed by anyone: two shapes can be similar without sharing the same foundational local elements (see pp. 90-93).
This argument of transposability is very useful since by demonstrating the existence of Gestalt quality, it also serves as a criterion to recognize a genuine one. Ehrenfels even proposes some kind of formalization of this recognition criterion: “If a presentation complex C is given in consciousness […] than that of the elements of C.” (p. 94; see the full passage) that could help the reader, he says.
One important point: regarding the title of §4 added by Smith, I am not sure that Ehrenfels actually brings any “proof” – in the sense of a pure argumentative or logical proof – but rather some “examples” which serve as some empirical (auditory) demonstrations. These demonstrations, in turn, can be used as exhibiting of the existence of Gestalt quality. If it’s a “proof” (and I am ready to admit it, if you push me on that point), perhaps it should be considered as an ostensive proof that takes the form of an intuitive evidence.
Please note, that contrary to mere trivial examples or far-fetched thought experiments that sound typically philosophical (I don’t say here that Ehrenfels doesn’t use any thought experiment at all), many examples come from his great musical experience, as a player and as a perceiver (see §4 in particular). The auditory fact of transposability comes from the scientist Ernst Mach with whom he shares some advanced musical expertise. Maybe my point is not entirely clear or understandable, but what I want to say here is that despite his very literary aspect, Ehrenfels’ approach appears to me highly empirical, if not experimental in some ways. Instead of examples coming from a purely philosophical imagination, Ehrenfels uses his own musical experience to more or less systematically investigate the ways of appearing of phenomena. In that sense, this text is far from being dusty. I would rather say that many philosophers of perception should try again to perceive, experience and attentively observe by themselves!
I am just wondering here how far this kind of approach is from visual demonstrations that will be used later by people like Wertheimer etc.
Concerning one of your preliminary remarks and your questions, Alistair, I would say the following:
*The impossibility of similarity relations:
You have probably in mind the beginning of §8. I am not so sure that Ehrenfels would not have recognized himself the relevance of the psychophysical procedure of judgment of similarities and differences. Even though this extract is awkward and confusing, what is at stake concerns mainly the status of relations and the way we grasp them. In this paragraph, Ehrenfels intends to show 1) that relations are indispensable for Gestalt qualities but that they cannot be confused with the last ones and 2) that Gestalt-quality cannot be reduced to a sum of relations. Actually the whole controversy about Gestalt-quality is about the nature of relations: do they belong to the founding contents or are they applied from outside by the perceiver? Are relations mediately or immediately grasped? Ehrenfels writes: “The melody can he HEARD, the square SEEN, not however any similarity and differences of its individual tones or two spatial determinations.” According to Ehrenfels, there cannot be any immediate sense-impression (Empfindung) of similarity because it presupposes an activity of comparison that cannot be given with the compared elements.
But no doubt that the above mentioned alternative is far from being clear-cut in this paragraph. Actually, Ehrenfels oscillates between speaking about relation in general and some specific kinds of relations, such as the one of similarity for instance. He starts by admitting that “the relation [in general I guess?], too, falls under the concept of Gestalt quality” (p. 101) but then “it will not do, however, to identify the relation [in general …?] with any of the Gestalt qualities”. So do relations identify with SOME specific Gestalt qualities? If so, which ones? How is that possible to consider, as Ehrenfels does, that relations are in some way indispensable for Gestalt qualities (and keeping in mind that “Gestalt qualities are given in consciousness (…) without any activity of mind specifically directed towards them”, p. 112) and that “it (relation in general??) cannot come into existence without some contribution on our part, without the specific activity of comparison” (p. 102)? The exact relationship between Gestalt quality and relations remains unclear to me.
* Does “Bewußtsein” equate with our actual understanding of the word “consciousness”?
You write: “what I can’t get a handle on (as a non-German speaker) is whether the connotations of “consciousness,” esp. those it bears in contemporary analytic philosophy of mind, are the right ones to grasp Ehrenfels’ implications here, or whether something much more deflated is meant.” Ehrenfels thinks from the framework of Brentano and his mereological analysis of consciousness. Given that current philosophers of mind have intensively worked these past years on Brentano’s philosophy (according to few of them, Brentano should be considered as a forerunner of the Higher-order theory of consciousness, see Uriah Kriegel for discussions), I would be inclined (regarding the very specific angle of your question) to respond positively. But maybe I don’t get everything you had in mind.
* “Relation”:
In German you have few words that can be loosely translated in English by “relation”: Relation, Beziehung, Verhältnis. That’s a thorny issue particularly because some of these German psychologists use these three words quite interchangeably. Conversely, Stumpf always uses the word “Verhältnis” when he speaks of fused tones. It means that there is an intrinsic and reciprocal connections between elements that are fused: that clearly cannot be translated by “relation” but maybe by “relationship”?? In French, one says “rapport”.
Actually the reference to Lotze is cosmetic or superficial regarding the direct and strong influence of Meinong on that very issue. Please note that Lotze is a mere second-hand reference extracted from Meinong’s key article “Hume-Studien II. Zur Relationstheorie” (1882). Meinong’s text is absolutely fundamental because it is responsible for the introduction of the British empirism in the Austrian tradition (if you except the deep influence of Hamilton and John Stuart Mill on Brentano). In the Brentanian tradition, if you want to understand the concept of relation you have to go back to the humean concept of relations of ideas and its interpretation by Meinong. For those who read the French, I recommend an article written by Ronan de Calan: “La caractéristique empiriste: La théorie de la relation de Hume à Ehrenfels.” (Les Etudes Philosophiques 2003/1, pp. 53-63à
* I totally agree with you concerning the extraordinary but unconvincing enlargement of the concept of Gestalt: it turns to be a catch-all word…
Excuse my belated joining of the discussion and please correct me if I’m wrong, but there seem to be two confusing kinds of infinity concerning the proliferation of Gestalts in Ehrenfels’ article. Based loosely on Hegel, the first kind could be called ‘bad’ infinity and the second ‘good’ infinity.
Ehrenfels discusses the ‘bad’ infinity on p. 88-9 in the context of possible objections to the existence of Gestalts. The objection concerning infinity goes as follows: if the elements e1 and e2 constitute (the Gestalt quality) e3, then e1 and e3 must also constitute e4, e2 and e3 must also constitute e5, ad infinitum. Supra-summation thus supposedly involves an infinite proliferation of entities. Ehrenfels argues against this objected ‘bad’ infinity by saying that this kind of infinite proliferation occurs only on a conceptual and logical level, whereas the ontological region in which Gestalts exist is the realm of perceptibility. In perception, however, the notion of infinite proliferation is inappropriate or at least unproblematic, because if we e.g. divide a colored surface into smaller parts and these parts into even smaller parts, we indeed perceive ‘infinite complexities’, as even the smallest colored part possesses more than one and thus indeterminably many spatial determinations. But the colored surface still does not proliferate infinitely in front of our eyes, i.e. reality itself does not change despite of the infinite combinability of elements. The same holds true for Gestalt-perception: we can perceive e1, e2 and thus e3 (as Gestalt quality), but the whole perceived entity (e1+e2+e3) does not expand ad inifitum in perception, even if we deliberately combine e.g. e1 and e3 and ‘see’ or ‘hear’ e4. The conceptual laws of logic and its limitations do not seem to hold true or are not problematic here.
Interestingly, in his 3rd Logical Investigation, Husserl picks up this problem. He suggests that on the logical-conceptual level of a formal ontology, it is only independent and separable parts (pieces) that can found new entities, but unlike Gestalts, these new entities are just a fusion of the pieces and not independent (‘transposable’) or even ontologically prior to what they are composed of. An ontological entity would thus be e1 + e2 = e3 (summation). In the case of dependent parts (moments) on the formal-ontological level, there is no infinite proliferation, because these parts are continuous and therefore in principle not distinguishable such that their part-character could come into play and be a foundation for any Gestalt-quality. Only dependent parts (moments) on the level of a material ontology of perception can create such transposable, independent wholes (Gestalten), but then again, like in Ehrenfels, we do not perceive reality itself or the contents of our perceptions to proliferate infinitely. In fact, the proliferation stops at a certain point. As Ehrenfels writes: “For a given complex of presentational contents given in consciousness only those Gestalt qualities are present whose foundations stand out noticeably from their surroundings.” Other theoretically possible Gestalt qualities are not present, although logically they would lead to a ‘bad’ infinity of the entity in question.
On the other hand, there seem to be a ‘good’ infinity, which comes into play in the notion of ‘higher order’ Gestalts. This has already been discussed in this thread and I just refer to the previous comments and §9 of the original text. This infinity is ‘good’, because it guarantees creativity, i.e. the origin of something new. What is more, the possibility of this infinity seems to account for the development of basically everything new, not only in art, but both in “nature” and “human creation” (p. 106) and in “more than half of all the concepts employed in everyday life.” (p. 108) However, if this second kind of infinity is also (partly) conceptual and not only perceptual, then I wonder where we should draw the demarcation line between the first and the second kind of infinity. Also, I wonder how we should ‘think of’ Gestalts, when on the one hand we are not supposed to think of them as infinitely proliferating due to their possible formal combinations with their underlying elements, but on the other hand we are supposed to think of them as being able to infinitely produce new Gestalts of a higher order when combined with other Gestalts. Finally, Ehrenfels rejects the objection of the ‘bad’ infinity by saying that not all presentational elements bring forth Gestalt qualities (p. 88), but he does not make such restrictions for the development of higher-order Gestalts, where there then should be – consequently – an even higher possible amount of infinite proliferation of Gestalts. Why is this higher potentiality of infinite proliferation not problematic then? And why and how does it lead to order and not chaos (here Ehrenfels’ later “Kosmogonie” becomes relevant)?
I’m not sure I have made my concerns sufficiently clear, but I’d be happy for any comments.