Reduction and Independence of psychology. A rapprochement

Personal research on the subject of reduction and independence of cognitive psychology

Performer: Orlin Baev, F32250, New Bulgarian University
Professor: Lilia Gurova, Ph.D.

Reduction and
Independence of
Higher-level Sciences
A Rapprochement

Can psychology be reduced to neuroscience? Let us imagine that neuroscience has progressed immensely and each mental state, phenomena and process has its direct neurological explanation. Than really the independence of psychology would be highly threatened – but how? It would be threatened as an explanation. For example, someone is now thinking about the love and its role in life, its manifestations on cognitive, thought level, as harmonized with the society and ones own personality logical process. If we suppose that at that very moment neuroscience has dived deep enough into the brain mechanisms, it will track exactly the zones involved and the consequences of their change, the biochemical composition in different neural networks, so on. But this very neuroscience, even though quite advanced at the imagined moment, would not be able to say a word about the exact cognitive content of the explored thinking process. So far the cognitive neuroscience, neurology and biochemistry are able to grasp to large extent the perceptual and emotional life in various cognitive biological systems, including the human cognition. Neuroscience might guess indirectly about which exact emotion is triggered at the moment, on the base of its knowledge about the connection between specific subcortical zones, neurotransmitters and certain emotional states. This is enormous achievement and it deepens every year with the advancement of neuroimaging and biochemical research. We certainly already know about our neural and chemical substrates of love, affection, anger, fear, aggression, mood’s ups and downs etc. And if cognition was bounded only on perceptual and emotional levels, neuroscience definitely would replace every so far existing psychological theory as already needless. But this lucubration is pure fiction, of course. If it was so, you, the reader of these rows would see on the screen only some unknown symbols. You would not even know these letters are symbols, but would see them in a way cat see them. If emotions were the peak of our cognitive abilities, no science would exist and no philosophical discussion will be pursued. Who knows, may be somewhere in the huge space there are different lines of evolution, not always obligatory involved in logical thinking… But we can not judge about that solely on imaginary premise. In the domain of cognition, we as a sapient race are too savage and cut off from the immense possibilities of cognitive evolution in the huge space. We can draw conclusions only from our own experience and objective knowledge as sensible human race. Such sane conclusion says that the top of our cognitive abilities is the attentive process plus working memory involved in logical verbal and serial or more intuitive, massive and parallel pictorial information processing – i.e. THINKING. Is neuroscience as it is now able of capturing one single thought? One single verbal, pictorial or mixed thought? How about series of thoughts, as they usually flow naturally? Even though neural imaging techniques might be capable of capturing the activity of prefrontal cortex spatially or temporaly, it is not good enough up to now for both. It will develop, of course and this disadvantage will be resolved and the preciseness for sure will become much greater, the devices much more useful and convenient. But even then, perhaps in at least 10 years from now on, will neuroimaging be so evolved to capture not only the neural correlates of thought, but thought itself? It might be possible, but it most sure requires qualitative brake through that so far is within science fiction. What we can say about memory – one of the main subjects of psychological research? Neuroscience advances and this is great. We already know a lot about hippocampus and its role, about the eventual main location of procedural memory – the cerebellum, the emotional memory – limbic system. Not much, but it is exiting continuous and very promising research. Regarding episodic and semantic memory things are more complicated, namely because it is involved in higher order cognitive process – thinking. Is neuroscience developed enough to research the cognitive content of memory – not only its hardware but its software, all stored episodes and their semantic sense? No, it is not and most likely will not be in the next few decades. Neuroscience can explore the spiking rates of the neurons, the exact kind and quantity of neural mediators acting in certain higher level cognitive process. But neuroscience is completely impotent in grasping even one single thought, mental image, or the way they emerge from neural and biochemical reactions. This emergence so far, from the viewpoint of the contemporary scientific knowledge is pretty much inexplicable and even mystic process. So far. What actually represents our thoughts and memories? May be our subjective reality, i.e. thoughts, memory, so on is solely computation based on the input sensory channels through lifespan. When an infant learns, it gradually enriches its internal representations of the objective outward reality, forms memories and slowly gains cognitive abilities such as analogy based associations of induction and deduction. We can say that brain, using visual, auditory and other sensory inputs and their storage (memories) in form of neural nets connections strengths just transforms them into subjective reality, implicitly perceived as a whole inner world of abstract ideas, rules, norms and understandings – believes. As I mentioned, cognitive science and neuroscience is not very clear about the exact way of transformation and thought emergence. It lacks the bridge links and processes. Of course, neuroscience is everything but stagnant – perhaps it will reveal these bridge processes soon. At least we can hypothesize so. Will this be the death of the psychological independence as a science on its own level with its own methods and terms? No! Why do I think so? Although Bechtel (whose book inspires me write this text) obviously is not enthusiastic supporter of the “Multiple realizability” hypothesis (as I am not very enthusiastic about mechanistic reduction), my opinion differs on that matter. Multiple realizability uses the term “realizer” to designate the mechanism underlying and producing some mental phenomena. It states that one and the same cognitive product can be realized by different, sometimes extremely different realizers. For example the cognitive phenomena pain (Bechtel’s example). In the various organisms such as mammals, reptiles, mollusks, etc. the pain can be produced by very different “realizer”, different neural/ brain mechanisms. Reptiles for instance do not have cortex and mollusks have only limited neural realizers – ganglia. But pain, the final product of these different realizers is still produced in each of these species. Numerous examples can be given. The very first time when I read about “multiple realizibility” I “misunderstood” the concept. I grasped it in reversed way – that one and the same neural/ brain zone or neural network can produce multiple mental phenomena, realizations. Now, having second thought on the subject I am continuing to believe it is important part of the “multiple realizability” conversation and its contributive role for the psychological independence. For example, the very same “realizer”, the amygdale can produce at least two opposite cognitive states: fear and anger. It might be involved in the pleasure and affection states as well. But such example is too crude. Let me go on higher cognitive level to defense my viewpoint. We can have one and the same underlying neural network with exactly the same biochemical reactions – but this same network can realize thousands of different cognitive products: thoughts. Of course, this is just general conclusion, because in each of the different thoughts we will still have small differences in the neurotransmitters and neural connections, hard to measure with the contemporary scientific means.
The “multiple realisability” thought direction in the scientific discourse is used to establish the independence of the higher level sciences, in our case – psychology. If many realizers can produce the same cognitive product than reduction and replacement of psychology by neuroscience is just “tunnel vision” of the eliminative reductionist scientists. Following this reasoning direction, we can say that each of the scientific fields has its own level and has to deal with the processes on that certain level. Neuroscience has its own terminology, research methods as well as psychology does. But this reasoning line has drawbacks – we miss 1) the bridge processes between higher and lower level and 2) the interdependence of both. My view point on “multiple realizability” enriches a bit the philosophical discourse. If we not only have many possible realizers (the classic view), but we assume that one and the same realizer can produce numerous cognitive realizations, this point strengthens psychological independence even more. How can we reduce psychology to neuroscience if we miss the huge variety of cognitive phenomena?

Regarding interdependence between lower and higher cognitive levels, we can consider neurological properties and formations such as brain and the processes within these formations as “low cognitive level”. “High cognitive level” then would be the subjective phenomenological experiences such as feelings, affections and of course, thoughts – pictorial and verbal ones. Considering high cognitive level we can not skip the unconscious, subliminal informational processing. But, just for the sake of convenience, let me focus for the moment on the conscious, traceable processes. How human cognition is interdependently interwoven with its neural correlates? Obviously these two are the very same – just parts of one whole system, operating on different levels. The one influences the other and likewise. Changes within our brain systems – neurotransmitters, brain parts so on have direct impact on the cognition. Likewise, conscious or unconscious changes in cognition triggered by the decision making process, metacognitive intention propelled by subliminal parallel cognitive work or by the environmental context – all of these cognitive “movements” on higher level have direct influence on their biological “hardware”, the brain, nervous system and biological correlates processing them. One successful comparison here is the computer metaphor. If one changes the parameters of hardware, software also changes its functioning – the speed of processing, memory storage, so on. Likewise, the software can have enormous impact on the hardware. Imagine you open your e-mail box and click on some interestingly looking message. If it contains worm or virus, it can damage or disable your hardware parts. Analogically in humans, if some maladaptive subliminal program (cognitive scheme) has infiltrated ones cognitive system (as a result of the early years experience, educational parenting style or other reasons), it can cause not only psychological mood changes or cognitive distortions, but even purely physical malady conditions such as headache, phantom pains, conversive disorder with bodily paralyses, psychosomatic diseases as asthma, gastric ulcer, diabetes and even cancer. As I said, these two cognitive levels: the low neurological and higher cognitive are intrinsically interconnected. None of them is more important than the other.

About the bridge connections between these two levels. One brave hypothesis might be that the bridge is realized on the subliminal, unconscious level, where cognition and its biological processor – the brain neural activities have a close rendezvous. In order this hypothesis to be developed, neither eliminative reductionism of theory reduction nor the more holistic and higher level approach of psychology need to be viewed on their own, but as different levels of functioning of one whole! This is my modest opinion.
But let us hear briefly what the traditional theory reduction perspective says.

Traditional theory reduction perspective

According to the theorists of the theory reduction perspective, it consists of:
Lower-level laws (in the basic, reducing science)
Bridge principles
Boundary conditions
Higher-level laws (in the secondary, reduced science).

Imagine we have complete advanced and definite explanation stemming from the extensive low level neuroscience research and we are fully capable of explanation of every and each cognitive higher level phenomenon in the terms of neuroscience. If it would be possible, higher level science such as psychology will at least loose its independence or even could be expendable. Of course, this speculation is far of the reality as it is. First of all, scientific approach does not have sufficient means to explore the bridge interlevel principles (so far). Secondly, even if we had them, the explanation would look like depicting some interesting story on the computer screen in some very basic computer language as assembler or in 1 and 0 –os. Even the programmers would need appropriate high level translation. And, as I mentioned above, even then we will have equal two way influence and interdependence of both levels. So, please, take a deep breath and slack – there is no way psychology to be deprived of its own level language, laws and independence as science. In my opinion, the low level science, the bridge principles, boundary conditions and high level science model is just perfect, solely the final stage: “Higher-level laws (in the secondary, reduced science).” has to be replaced with “higher level laws in the high level science”. This changes the whole understanding and transforms the reductionist model in reductionist-holistic, without the need one of them to renounce the other. And please, do not be scared of the word holistic. I am using it just to designate the two-way interdependence between low and high level cognition. In the chapter on which I am reasoning, Bechtel mentions that even if low level brain research reaches much of the actual bridge principles and boundary conditions, even then high level psychological approach would be needed as a heuristic for further low level neuroscience research and as a shorthand in the explanation.
Interesting and very valuable suggestion is given by McCauley:

As we see on the graphic above, McCauley suggests that theory evolution in some domain happens horizontally and vertically interdependently. When older theory on lower or higher level is being replaced, it happens with another theory on its own level – low or high. If older neuroscience theory is replaced, it will be with new one neural theory. If older one psychological theory is replaced, it will be from new one psychological theory. On whatever horizontal level change happens: low or high one, this change has direct vertical impact on other level theory and results in change as well. If psychological research reaches some new theory, it will give important clues for the lower level neuroscience research and will result in neural theory change as well. And likewise, if neuroscience achieves new theory, it will inevitably propel psychological theory change too. The levels and their development are interdependent, but each level has its own means and independent research: “the upper-level theory lays out regularities about a subset of the phenomena that the lower-level theory encompasses but for which it has neither the resources nor the motivation to highlight. That is the price of the lower-level theory’s generality and finer grain (McCauley, 1996, p. 31).”


Bechtel begins his mechanistic perspective introduction with the pure and clear stating his preferences of this viewpoint. Afterwards he clarifies the notion of levels. Firstly, are levels defined by the notion of size? No, they are not. Although it seems intuitively proper understanding, as a matter of fact size has nothing to do with levels. If all entities of same size form a level then snowman and a person would be at one level, or the society in some town had to be at the same level with field with rocks with the same size as a town. The size is inappropriate for levels differentiation also because if only the same size entities could form a level, different size entities couldn’t interact. But obviously they can. Elephant can interact with the grass and the louses can interact not only with the elephant skin, but with his mind too. So, which one is of bigger size – the mind or the louses? Apparently, the size is not the best criterion foe level or strata.
“Another proposal involves bringing together two different notions of level in construing phenomena, and considering how this maps onto disciplinary divisions (Abrahamsen, 1987). The mereological notion of level is subordinated to a different notion of levels in which phenomena are grouped in a way that generally corresponds to common academic divisions: ordinary physical phenomena (physical sciences), phenomena of life (biological sciences), behavioral/mental phenomena (behavioral sciences), and phenomena involving products of human behavior/thought (humanities and social sciences). Only within each of these four levels do mereological levels come into play: each has its own part–whole hierarchy that is unlike those at other levels.” This is really interesting proposal and in my own opinion, the closest one to the actual picture of reality that we are trying to represent philosophically. Bechtel infers that “each higher domain is generated from the lower one; entities and processes in the lower domain tend towards increasing complexity, and organized systems emerge for which specialized concepts and explanations appear necessary.” I can say that this is partial view point. If the lower domain entities generate and influence the higher ones, the higher domain entities influence the lower ones reciprocally. Let we have one example: the citizen’s society. If the separate entities/ persons of this society get sick from some pandemy, the whole society will be influenced on higher level, as social organism. Likewise, if the whole social organism on higher level, represented by, let say, the president, take inappropriate decision, the many single entities/ persons will be influenced. Anyway. “The mechanistic account offers no way to evaluate whether the components of a mechanism are at the same level as entities outside the mechanism. For those seeking a global account of levels, this may seem to be a serious limitation. Yet, as I just noted, that view of levels is problematic. Local identification of levels is sufficient for understanding levels in a mechanism and for capturing how mechanistic explanation is reductionist.” Yes, mechanistic account helps, perhaps namely because of its over simplification. It allows the scientists to perform their research having sufficiently good theoretical frame. Not the exact one, but useful one. Bechtel cites two views on emergenism that deserve our attention and conscientious critics: “Damn near everything we know about the world suggests that unimaginably complicated to-ings and fro-ings of bits and pieces at the extreme microlevel manage somehow to converge on stable macrolevel properties. On the other hand, the ‘‘somehow” really is entirely mysterious … (Fodor, 1997, pp. 160–161).”
“The bigger mystery is why this seems at all mysterious. Task analysis of two significantly
different realizations of the kind corkscrew will reveal how it is that each realization is able to remove corks. Once one understands the principles by which waiter’s corkscrews remove corks, and the principles by which double-level corkscrews remove corks, what is so puzzling about how two causally distinct devices
can “do the same thing”? (Shapiro, 2004, pp. 161–162)”

The subject of emergence is bounded with that one of reduction – independence, because of bottom up interlevel causality of emergence. How exactly the mechanistic parts form all these purposeful functioning wholes? If we have for example the machines, programmed by men, the question who gave certain aim assembling the parts is clear. Regarding biological cognitive systems we have only very vague theories such as Darwinian one, often obviously questionable. But is is hazardous that supporters of mechanistic view have taken for example exactly the mechanisms? No, it is not. The mechanistic account is so far very limited, narrow and does not take into account the whole picture in the nature as it is, on any level, between levels and out of certain mechanism and its context influence. That’s why Shapiro gives for example corkscrews, but not human conscious high level cognitive processes. But, to answer to my own critics, the AI research has some role in this discussion: the artificial neural networks exercise completely new behavior and even can learn independently, i.e. new one higher level processes emerge from lower level ones. In reality, the cognitive levels are many. Mechanistic approach sees them as two. Mechanistic reduction links bottom up and top down reactions without causal explanations involved, but just mechanistic reactions. Well, convenient view point, but limited and reflecting the limited understanding of its supporters. Mechanistic point of view does not take into account the multiple realizability fact as well. Namely because it is mechanistic and by definition extremely cut off the whole picture point of view.

My conclusion is as follows. Neither theory reduction perspective, nor mechanistic one represent the things as they are. Perhaps one realistic perspective, wider and encompassing the whole picture is needed to come into the stage! For now, the most appropriate view in my opinion is that one:

Lower-level laws (in the basic, reducing science)
Bridge principles
Boundary conditions
Higher-level laws (in the secondary, high level / instead of low level science as stated originally/ science).

Note: My thoughts are based on the chapter four “ Reduction and Independence of Higher-level Sciences-A Rapprochement “ of William Bechtel’s book “Mental Mechanisms: philosophical perspectives on cognitive neuroscience”.

Orlin Baev,
cognitive psychologist and cognitive psychotherapist

Вашият коментар