Volume 24, Issue 2 (summer 2022)                   Advances in Cognitive Sciences 2022, 24(2): 156-167 | Back to browse issues page


XML Persian Abstract Print


Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:

Rezayati Charan A, Gharibzadeh S. How realistic are we?: Investigating the degree of alignment between our perception and the objective reality. Advances in Cognitive Sciences. 2022; 24 (2) :156-167
URL: http://icssjournal.ir/article-1-1411-en.html
1- PhD Candidate of Cognitive Science, Institute for Cognitive and Brain Sciences, Shahid Beheshti University, Tehran, Iran
2- Associate Professor of Cognitive Science, Institute for Cognitive and Brain Sciences, Shahid Beheshti University, Tehran, Iran
Abstract:   (331 Views)
Introduction
Opinions about perception regarding its relation to reality fall into a spectrum in which, at one end, it is believed that our perception is entirely in line with external reality, which is called direct realism or naive realism. At the other end of the spectrum are those who believe there is no correspondence between perception and external reality and are so-called anti-realists. Classical theories of cognitive sciences believe that the perceptual system can guarantee the survival of an organism when it represents external reality as accurately as possible.
Nevertheless, some theories like the Interface Theory of Perception belonging to Hoffman et al. believe that our perception has nothing to do with external reality. According to this theory, the direction of evolution in shaping the perceptual system is not a representation of reality but a guarantee of the survival of living beings with the least energy; they show with these assumptions that perception has nothing to do with reality.
However, they have neglected the assumption of environmental changes in their model. In our previous study, we modified their model and presented a new one considering environmental changes' assumptions.  This research showed that the species perception approximates external reality, with a considerable difference in having a better chance of survival than other species. The present study intends to investigate how well this approximation can be in terms of its proximity to reality to maintain the existing survival.
Methods
The method used in this research is agent-based modeling. Based on the conceptual model proposed by Hoffman, our model has three types of agents: the truth agent, the simple agent, and the interface agent. Each agents is generated in the desired number at the beginning of the simulation and randomly distributed with a uniform distribution in the simulation environment. Each agent is assigned an initial energy value based on a normal distribution. Each patch of the environment is also assigned a random amount based on a normal distribution function, such as the feed capacity of that patch. Each agent evaluates its neighbor's patches based on its perception function and selects and consumes the most desirable patch resource. Then, at the end of each step, the gain of each agent is calculated based on the following function:
Gain=Ux-C
Where U (x) is a function of the utility of each agent, which is in accordance with the model proposed by Hoffman et al. as follows:
Ux=ae-
The cost function or C is also defined as follows:
C=celog2q+ckqnb
In this modified model, the possibility of changes in the environment is also considered. In case of environmental changes, the species with the interface perception loses the correspondence between its perceptual interface and the environment and makes a mistake in distinguishing the desired and undesirable area., the energy function of the interface agent in step n of the simulation, taking into account the assumption of environmental changes, is calculated as follows:
Eninterface=En-1interface+
The investigation conducted in this study is to determine the maximum accuracy that the simple species (species with an approximation of reality) can have in its perception of the environment. The model used in this study is a modified version of the model made by the authors in their previous work that was introduced earlier. Here, instead of assigning a fixed value to the number of perceptual categories for the simple species, the number of these categories changes in each simulation run. The efficiency of each perceptual strategy in these different simple species situations is examined.
Results
Each combination of different values for different model variables such as cost per bit of information, resource growth rate, desired area boundaries, probability of change of environment, the initial number of factors, and the number of perceptual categories of approximating factors is repeated 20 times and the average of these repetitions, as a result of the simulation with those specific settings is reported. The main result of the simulation is demonstrated in the following figure, which shows the performance of each species concerning different numbers of simple agents' perceptual categories.

Figure 1. The final population of different species concerning simple agent's different precision of perception

The results show the efficiency of different perceptual strategies concerning changes in the number of perceptual categories of the approximate strategy, the probability of environmental changes, the cost per bit of information, and the growth rate of food resources. For all these parameters and in almost all parametric spaces, the simple strategy, with an accuracy of 40% or less than the truth strategy’s accuracy, shows better performance than both the interface and the truth strategy.
Another part of the results reveals the performance of each species for different values of variables of the probability of environmental changes and cost per bit of information. The results show that in almost the entire parametric space, the approximating species perform better than the other species in both cases. Thus, it can be argued that in environments with extended periods of stability or in isolation from the general environment, perceptual interface evolvement and severe disconnection between perception and reality can be possible. However, in dynamic environments, even with relatively low dynamics, survival requires that the perception system provide an approximation of reality.
The results also show that with increasing approximation accuracy in the simple species (those with an approximation of reality), even up to 50% of the realistic species, if the cost per bit of information is low, the simple species still performs significantly better than the species with the interface and at high values of cost per bit of information in some parts of the parametric space still performs better than the interface type. In the other parts of the parametric space, it works at least as well as the interface type.
Conclusion
According to the present study’s findings, evolutionary mechanisms allow the perceptual apparatus to be accurate, to about half the accuracy of an accurate realistic perceptual apparatus. Indeed, the obtained results show that not only does evolution not lead to the formation of perceptual systems unrelated to reality, but it also allows for a much more accurate approximation of perception of reality to a considerable extent. So we as evolved beings can hope that what we perceive is reasonably similar to what exists in the real world. Nevertheless, we should keep in mind that the circumstances under which the natural selection mechanism achieves such a level of accuracy in perception may also depend on other factors. For future studies in this area, the first step is to find a ground-based criterion to examine the accuracy of perception, which is reasonably easy to calculate and measure, and applicable in a variety of cases. In the next step, finding experimental contexts to test the hypotheses is very helpful. Seemingly, this research should pay much attention to environmental and ecological features and study the perceptual system of native ecosystems in a range of ecosystems with different levels of diversity and history of change.
Ethical considerations
Compliance with ethical guidelines
This study is based on mathematical models and their simulations and does not use experimental data from human or animal studies. So, no subject requires compliance with the code of ethics.
Authors' contributions
The authors have equal contributions to forming the main idea of this study. Shahriar Gharibzadeh supervised the research. Arman Rezayati did modeling and simulation and also wrote the paper.
Funding
No financial support was received from any institution in conducting this study.
Acknowledgments
The authors did not use anyone else's help in doing this study.
Conflict of interest
There was no conflict of interest between individuals in this study.
 
Full-Text [PDF 1152 kb]   (45 Downloads)    
Type of Study: Research |
Received: 2022/04/12 | Accepted: 2022/05/19 | Published: 2022/08/11

Add your comments about this article : Your username or Email:
CAPTCHA

Send email to the article author


Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Designed & Developed by : Yektaweb