Typically, we look at the world, in our house we see things that are there, we left to the street, we found various things, we see cars, we have entered a pastry shop to buy chocolate.
Certain type of nerve cell activation is co-occurring with our look at the world (or with our sense and experiencing the world). In particular, certain types of nerve cell activation typically co-occur in our visual cortex.
However, presumably there are so many ways the nerve cells of the visual cortex are activated by the things we see, as the properties exemplify by the things we see.
The so-called binding problem is the problem of knowing how nerve cells (extended to other areas than the visual cortex) co-occurring with the stimulation by things out there are activated, what is the configuration of this nervous activation.
However, the "binding" of this problem is binding, for example, of the car with the street of Lisbon (environmental binding, for example, Farah 1990), is binding of color, shape, luminance and movement of the car (binding characteristics, for example, Gulyas and Roland 1991) and is the binding of what characterizes each of these characteristics (binding characteristics of characteristics).
The binding of what characterizes each of these characteristics, for example:
it is binding of the color with frequencies and wavelengths: the blue of the car with frequency 3801-668 terahertz and 380-450 nanometer wave ( binding of frequency), for example, Graham (1969);
it is binding of the form with the curved and straight lines the car happen to exemplify (binding shape-curves), for example, Cavanagh (1967);
it is binding of the movement with the speed of the car (binding movement-speed), for example, Stoner and Albright (1993);
it is binding of the things with intervals of time (temporal binding), for example, von der Malsburg (1981),
and finally
it is binding of the things with your location (location binding), for example, Ungerleider and Haxby (1994).
As the use of concepts like the concepts of "cognitive" and "cognition", the use of concepts like the concepts of "binding", if it neglects the phenomenology, is improper and exaggerated and may be preventing progress in the search for new treatments and medications for psychiatric disorders and neurobehavioral disorders.
Concepts like the concepts of "binding" has been an umbrella (erroneously perceptually imagine instead of empathetically imagine), but its use has been now revised (Di Lollo 2012).
Note that the best way to review statistical results (probability of co-occurrence of cerebral physical, chemical and functional characteristics with conscious mental states) is to replicate these results and it is currently large the discussion in the literature, for example, Open Science Framework (http://openscienceframework.org), The PubPeer Foundation (https://pubpeer.com/recent) or Curate Science (http://curatescience.org).
The suggestion of authors such as Ioannidis (2005) and Moonesinghe et al. (2007) is that p values null hypothesis significance test (NHST) less than 0.05 (type I error , alpha, rejecting the null hypothesis when it is true) given the hypothesis H that two samples Y and Z do not differ or have no effect on each other (the p value as typically occurs in statistical research about the phenomenological, physical, chemical and functional characteristics co-occurring with consciousness mental states), although acceptable for publication in peer-reviewed double blind journals, cannot precluding results that appear to be positive, but which are in fact false positives.
A p value less than 0.05 does not necessarily mean that an effect is out there, in the world. This means that, for a 95% confidence interval, there is a 5% probability that we will reject a true null hypothesis (inversely propositional to type II error, beta, failing to reject the null when the alternate hypothesis is true). It will be apparently good fortune, particularly if a laboratory study is an underpowered study. A p value large than 0.05 implies that the data could easily have been observed out there (in the world) under the null hypothesis. However the data could also have been observed under a range of alternative hypothesis rather than the binary framework of the null hypothesis significance test.
Other authors, such as Carp (2012), discuss what he calls the "secret life" of the methods reported in the functional magnetic resonance imaging (fMRI) literature or the plurality of methodological worlds in the fMRI literature.
Typically, as the authors (for example, Sober and Kording 2012) working on empirical research about the brain explicitly recognize, the results raise more questions than they answer to the questions with which they started: “However, at some level the paper raises as many questions as it answers, since there seem to be two completely independent (but not mutually exclusive) interpretations of the results”.
Nerve cell configurations co-occurring with stimulations for things out there do not represent those things: in chemical reactions and brain electrical stimulations there is nothing like the propositional content to relate sensations and experiences (and mental states like the phenomenological elements of attitudes) as there are to relate mental states as beliefs, desires and intentions with things out there.
The idea that neurons (your physics and chemistry) "encode" or represent "information" is wrong.
The idea that brain physical, chemical and functional characteristics process information is not empirically established by cognitive neuroscience, it is a metaphor (erroneously perceptually imagine instead of empathetically imagine). The concepts of "processing" and "information" are concepts of popular psychology that seem scientifically rigorous, but are not scientifically rigorous (for details, for example, Burock 2010 and Alivisatos et al. 2012).
These concepts of "processing" and "information" (like the concepts of "cognitive", "cognition", " binding" and, as we shall see in a moment, like the concept of "attention") help researchers to communicate the allegedly results of their research (only they have been introduced by the apparent facility of communication, but they are fallacious since brain physical, chemical and functional characteristics are brain electrical stimulations and brain chemical reactions), but they are not one of these empirical results. If brain physical, chemical and functional characteristics “encode” or “represent”, it is to assume from the start what you need to show, if brain physical, chemical and functional characteristics de facto “encode” or “represent”. What is here is a petition of principle.
For example, apparently it is more easier to say that the brain area called the fusiform gyrus (temporal lobe) processes human facial information (data interpretation) than to say that certain physical, chemical and functional characteristics of nerve cells configurations (empirical results) are probably co-occurring with the presentation of images of human faces.
In the apparently relevant literature, the fusiform gyrus comes to be called “the fusiform face area" when the presentation of the image of the human face is to be co-occurring with nerve stimulation of the cells in brain fusiform gyrus (temporal lobe) area.
The area wrongly named this way, “the fusiform face area", is also stimulated by the presentation of images of familiar objects that are not human faces.
What stimulates the cells in brain fusiform gyrus (temporal lobe) area is out there, it is enough the context of high probability of the appearance of the image of the face, the image of the human face may not even be outside (Cox et al. 2004).
The image of the human face is a necessary but not a sufficient condition for co-occurring nerve stimulation of the cells in brain fusiform gyrus (temporal lobe) area. If I say, as sometimes, we hear, that “the brain fusiform gyrus (temporal lobe) area is necessary” for recognition of an image as human face, this is not because of images of human faces (they might not even be outside) but because the fusiform gyrus (temporal lobe) area, as all other brain's areas, are necessary for a healthy brain. As, for example, Dronkers et al. (2007) remind us, the Broca's aphasia is caused by wider brain injury area than the Broca's area: cerebral cortex, the so-called white matter, the insula, basal gland and portions of the anterior superior temporal gyrus (see, above, p. 14).
What there is is a certain probability of neural physical and chemical reactions occur under certain laboratory conditions of stimulation. These laboratory conditions are simulations of things out there, there are no things out there. Images of human faces, are not human faces. In a typical experimental design there are images of things out there, there are no things out there.
One thing is the probability of the evidence given the hypothesis. Another thing is the probability of the hypothesis given the evidence.
By Bayes theorem, where "p" stand for the probability, "hpa" stand for the alternative hypothesis, "hpn" stand for the null hypothesis and "e" stand for the evidence,
p (hpa/e)/p (hpn/e) = p(e/hpa)/p(e/hpn) . p(hpa)/p(hpn)
the a posteriori probability
p(hpa/e)/ p(hpn/e)
is the multiplication of the a priori probability
p(hpa)/p(hpn)
by
p(e/hpa)/p(e/hpn),
a quantity known as Bayes factor.
The computation of frequency factors for ANOVA models p(e/hpn), in contrast to the computation of Bayesian factors p(e/hpa)/p(e/hpn) for ANOVA models, has been available in common and friendly programs.
The computation of Bayesian factors for ANOVA models has not been available in common and friendly programs, but it is beginning to be (Rouder et al. 2012, Sun et al. 2012 and JASP, a low fat alternative to SPSS, a delicious alternative to R. Bayesian statistics made accessible, from 2014 until now JASP team released the 8th version 0.8.1.0, March, 2017, https://jasp-stats.org/).
For example, instead of the probability that empirical results are in this or that way p(e/hpa)/p(e/hpn) . p(hpa)/p(hpn) (given our hypotheses related to behavioral and brain imaging results) - because we use frequentist statistics we are not going to make everything seem frequentist - my suggestion (assuming bayesian statistics) is to update degrees of belief in our hypotheses p(hpa/ e)/p(hpn/e) (given the behavioral and brain imaging evidence).
The physics and microbiology of today is a far from what, for example, Lucretius (who died about 50 BCE) said in his De rerum natura (for example, in Books II and IV) about light and microbes but about attention today we continue to say qua Lucretius that we only see clearly those things to which the mind has already prepared us, “save those for which the mind has prepared itself" (IV, v. 803-804).
Qua Lucretius today we continue to ask "do you not see that our eyes, when they begin to look at something that is tenuous, make themselves intent [contendere] and ready, and that, unless they do this, it is not possible for us to see clearly [cernere acute]?" (IV, v. 808-810).
As for Lucretius also for us "even in things que are plainly visible you can note that if you do not direct the mind [advertas animum], the things are, so to speak, far removed and remote for the whole time" (IV, v. 811-813).
Esterman and Yanti (2010): “selective visual attention directed to a location (even in the absence of a stimulus) increases activity in the corresponding regions of visual cortex and enhances the speed and accuracy of target perception”.
For example, “some discriminations appear to be made automatically, without attention and spatially in parallel across the visual field. Other visual operations require focused attention and can be performed only serially” (Treisman 1985).
Something explains the empirical differences, the p values null hypothesis significance test (NHST) less than 0.05, p(e/hpn), the terms underlined (by us) can typically be deleted (or replaced by "something") but the empirical results, for example, by Treisman (1985) are not affected at all: the inclusion of such terms as "without attention" or "require focused attention" they seem to be a cognitive gain, but they are not a cognitive gain.
Some definitions of "attention" include concepts like the concepts of "filter" (Broadbent 1958), "spotlight" (Posner et al. 1980), "magnifying lens" (Eriksen and St. James 1986), "glue" (Treisman and Gelade 1980), "competitive trend behaviorally relevant" (Desimone and Duncan 1995), but also these concepts can be typically eliminated that the related empirical results are not affected.
There is no such thing as attention (Anderson 2011). There are false dichotomies (Anderson 2011) such as the following:
top-down/bottom-up,
exogenous/endogenous,
split/focused,
feature/object,
local/global,
automatic/controlled,
voluntary/involuntary,
non-conscious/aware,
peripheral/central,
intraperceptual/extrapeceptual,
passive/active,
pre-attentive/attentive.
And now we add to the list of umbrella concepts like the concepts of "cognitive", "cognition", "binding" and "attention", the concept of "semantic" .
Semantics is not in the head, is out there.
There are co-occurring cellular configurations, for example, in particular (the details in Binder et al. 2009) the following "three general categories of cortical regions" are involved in semantic "processing": multiple posterior and heteromodal associative cortex, heteromodal prefrontal cortex, medial limbic regions.
For these authors, as Binder et al. (2009), in contrast to the brains of non-human primates, the expansion of these areas in the human brain may explain uniquely human abilities as productive use of language, planning, problem solving, the ability to create cultural and technological artifacts: for them, all of these capabilities depend on what they claim to be the fluid retrieval and efficient manipulation of semantic knowledge.
The question of where is the semantic system (Binder et al. 2009) it is not about semantics, it is about co-occurring nerve cell configurations.
My suggestion is that those studies may well use concepts like the concept of "semantics" as an umbrella (erroneously perceptually imagine instead of empathetically imagine) that, if eliminated, not affect the related empirical results.
Semantics is out there, certain things out there stimulate nerve cells, but the co-occurring configuration of these with that stimulation, if asserted "representational” or “informational or “coding", is just an undue and exaggerated use of concepts: neurons, their synapses, neurotransmitters, molecular receptors are cellular organisms.
This way of speaking is derived from linguistic categories.
Note that, expressions like "neural code" are not neurons, are us talking about neurons.
Neurons are to be things out there, neurons are being represented by us, but neurons themselves are not representations.
The statistical analysis of both the data invoked by those who argue that phenomenology can be separated from its functional characteristics (say, phenomenologists results) and the data invoked by those who argue that phenomenology cannot be separated from its functional characteristics (say, accessists results) is typically the p values null hypothesis significance test (NHST) less than 0.05. The empirical results in both cases may be false positives.
However, allegedly there are alternatives to the p values null hypothesis significance test (NHST) less than 0.05 (Killeen 2005), the Bayes theorem.
Notwithstanding, any of these phenomenologists or accessists results “was unlikely to be replicated when it came to most of the “other approaches” with which some would supplement or replace significance tests– notably Bayesian updating, Bayes factors, or likelihood ratios (confidence intervals are dual to hypotheses tests)” (Mayo 2017).
Thank you very much for reading, I hope to publish the remaining posts on the sixth day after the previous post (one post every sixth day, more or less).