5: Interpretation, Memory, Experience, Learning
The human mind optimizes itself to give greatest attention to what is new or different in the environment. We cannot focus on everything we perceive, or we would be overwhelmed - we must focus on perceptions that are important. As such, the familiar and the known are filtered out and treated as unimportant - and focus is given to what is unfamiliar, new, or different.
Memory, which may function as recollection, experience, or learning in various situations, is the tool by which we filter - to recognize something as a threat, an opportunity, or inconsequential. And while the effect of new stimuli seems more interesting and prescient, to overlook the importance of memory is a grave error in considering human behavior.
Why The Memory System Is Important
The majority of the brain is a warehouse of memories. (EN: The myth that "we use only 10% of our brain" arose from studies of activity, which focused only on processing immediate stimuli, and found that only about 10% of the neurons are active at any given time. This observation was misinterpreted to dismiss parts of the brain that are not engaged in the present task - they are not always active, but are not to be dismissed as useless. Said another way, we may use only 10% of the brain at any given time, but it is not the same 10% in use at all times.)
Neurologists have sought to identify a specific area of the brain devoted to memory storage, but have failed to do so. While the artificial logic, in the form of a computer, contains memory in a specific component (the hard drive), the human brain has many separate locations in which memory, both short- and long-term, is stored.
(EN: As such, a better analogy for the brain would not be a single computer, but a network of computers, each with its own processing and storage facilities.)
Also in terms on neurology, it is presently believed that the brain of every person is made up of the same structures. Everyone has the same structure of lobes and the same regions within each lobe, such that a specific part of the brain processes visual stimulus, and if two people suffer the exact same damage to the exact same part of the brain, the effects will be similar.
What differs among people is the way in which the various lobes and regions are engaged in a given situation. There's a great deal of overlap, but significant differentiation in what parts of the brain become active in a given situation. This is reckoned to be due to differences in the ways that the stimuli are recognized and processed, which come from differences in the way the brain is used, which comes from differences in individual experiences.
So in this way, experience and memory creates differences in brain activity: everyone is issued the exact same organ, but everyone uses it in an idiosyncratic way.
Feeling Good, Culture, Personality And Memories
In terms of brand, the most significant memories are feelings, rather than the raw interpretation of sense data. If we can create a positive "feeling" about a brand, we set the stage for a positive reaction to more factual information.
And in terms of neurology, feelings are processed in very specific regions of the brain - which gets back to the notion of differences in brain activity. When two different people sense the same thing, such as smelling a spicy food, one person's "fear" center may be activated whereas another person's "attraction" center may be activated.
Food is specific to culture. The author gives the example of "bitlong," a dish that is repulsive to most people (it's slightly rotten meat), but highly appealing to South Africans. The difference is that people of the South African culture are accustomed to eating it, it brings back comfortable memories of their homeland, and as such their response even on a neurological level is different.
This is true even when stimuli are imagined: the thought of public speaking, in the absence of physical stimulus, causes different people to experience different emotions based on their experience. A person who has been trained and has experience at public speaking will be comfortable at the prospect, perhaps even a bit excited to use their skills; a person who has not been trained and has been embarrassed in the past will be fearful.
This difference confounds researchers who believe that the physiology alone will explain human behavior. While it is true that certain parts of the brain are "hard wired" across all individuals, much of the "software" is unique to an individual.
The Brain
The human brain is astoundingly complex: it consists of about 10 billion neurons, and given that each neuron may be connected to several thousand other neurons, this leads to the estimate that the human brain has about 10 quadrillion synapses.
Even if each synapse consisted of a binary "bit" of information with an on/off value, as transistors operate within a computer, this would entail an astounding amount of data storage and processing capacity: but each synapse has several receptors where an axon can communicate data to a dendrite.
In all, the brain is a n astoundingly complex organ that science has yet to adequately analyze. Even the theories and models of current science are insufficient to grasp its full complexity, and continued research results even greater complexity than is already known. For example, it is currently theorized that axon-to-dendrite is not the sole method by which neurons transmit and receive data (axon-to axon, dendrite-to-dendrite, cell body to axon or dendrite), which adds yet another layer of complexity.
Neurons
Neurons are individual nerve cells that run throughout the body. Outside of the brain, they transmit signals to various organs and tissues that cause them to react in certain ways - the simple twitch of a fingertip results from several signals that cause a series muscles to expand or contract.
Neurons are connected to one another by a system of axons and dendrites - transmitters and receivers. A single neuron may have several thousand of each.
Communication is electrochemical. In a resting stage, a neuron caries a charge of less than 40 millivolts, which can increase by up to 70 when it is activated. These impulses travel at 100 miles per second, which is rather slower than the speed of electronic impulses through wired circuits (about 124,000 miles per second). After transmitting, the neuron returns to its resting state. A neuron can fire thousands of time per second, which is again much slower than electronic circuitry.
The result of this speed can be seen when people process sense data. It is suggested that a person recognizes an image of an animal within a tenth of a second, and also forms a memory of the image in the same amount of time. In that way, people recognize things that they have seen before, or identify it as being similar to what they have seen before, in tiny fractions of a second.
It is suggested that the process may be even faster than that, as we are limited in the precision of the equipment we use to measure brain activity.
Synapses
The synapses are areas of proximity between axons and dendrites, and are theorized to be the primary method of communication among neurons by releasing chemicals. But again, this is not the simple 1/0 binary of a transistor, but a more complex connection in which there are multiple receptors and transmitters, each of which may be sensitive to degrees.
(EN: As an illustration, a single synapse with three receptors, each sensitive to four degrees, can transmit about 144 different values. Most synapses have several receptors, and the degrees to which they are sensitive have not been estimated.)
Also, synapses and neurons do not fire independently. A neuron may fire thousands of transmitters at a given time. And through a process called "recruiting," two ore more neurons may fire at the same time.
It's noted that the electroencephalograph (EEG) is the device used to measure electrical activity - but even the most advanced EEG does so in a primitive way: it cannot measure a single transmitter/receptor, nor even a single synapse. Instead, the EEG produces a general reading indicating that there is activity in a given area, which may consist of thousands or millions of neurons firing simultaneously.
As such, we depend on the EEG as the most precise instrument at our disposal, but must acknowledge that it is vague and imprecise. It's the best we can do, at the present time.
Neural Networks
The physical properties of neurons and synapses, individually, is foundational science, but of less significance to the way that they perform as a system, in the context of a neural network.
Unfortunately, this is where hard science stops: the best technology we have today is incapable of precise measurement. As refined as it has become, we can only observe brain activity in a vague and general way, to detect when there is electrical activity above a certain threshold and a loosely defined "area" of the brain. Too much is happening too quickly for us to be able to accurately measure and observe it all.
So at this point, we return to observation and theory to infer what the patterns of activity - their intensity and location within the organ - actually means. That is, we can infer that, because a certain part of the brain is active when a person sees something, that the activity relates to the visual stimulus of that particular thing, but we do not have evidence of precisely which neurons are firing, precisely which synapses are transmitting or receiving - merely that a given area is active in a given test subject.
And as such, we can presently proceed based on these vague observations, amalgamated, to make an educated guess as to what is happening in the brain, in general terms. It's a much more informed guess, based on more detailed information of the science of the past, but on far less detailed information than science will provide in future.
It's mentioned that scientists are currently using artificial neural networks to study the function of the brain. While there is much to be learned from creating a computer simulation of neural networks, it comes with the caveat that what is being "discovered" is only as good as the theories that are implicit in the development of the model.
Gestalts
The term "gestalt" is used in psychology for a phenomenon whose "whole" is not perceived, even when we account all the various components we are able to consider. In terms of brain science, it considers that our perception of things in the environment begins with sense data, but becomes more complicated as it is processed by the brain.
A good example is the written word: a person who recognizes a word in writing first begins by analyzing its properties. Setting aside font, color, and context, they see a series of shapes, recognize them as letters, recognize the combination of letters as a word, and associates to that word a given meaning.
But this is only the beginning of the process, as even when the word is recognized, it becomes associated to a concept, to other words, and to other memories. The word "snake" conjures up a different visual image, different emotions, and different memories, for each person who reads it. And more, you have very little control over what the word causes you to recollect - the "snake" gestalt is assembled automatically rather than a conscious process.
An interesting parlor-trick the author suggests is to tell someone "do not think of an elephant." It cannot be done, because the moment the word is mentioned, it triggers the process of recognition and identification and brings to mind the gestalt of an elephant. The only way to not think of something is to be focused on something else, but at the very mention of the thing you are trying not to think about, you will find yourself thinking about it.
Another parlor-trick is to mention something a person may be unaware of - for example, the drone of an air-conditioning unit. The sound was there all along, and they were likely perceiving it, but they weren't conscious of it until it is mentioned. And after it is mentioned, they are "suddenly" aware of it, and remain aware of it until their mind becomes so engrossed with other things that it again fades into the background.
The last diversion deals with troubled sleep - how a person who wants to fall asleep finds themselves annoyed by things in their environment that would be unnoticed on any other night: the motion of air across their skin, being too warm beneath the sheet, and the subtle sounds from outside. There are many sensations that become a distractions, preventing them from falling asleep, and they must focus on something else to ignore these nuisances - the hackneyed concept of "counting sheep" in one's mind as a way to fall asleep.
The author describes the human mind as "a pool of developing gestalts." Everything we focus our attention on, and possibly quite a few things we perceive without consciously giving attention, contributes to a gestalt - the perception is received, categorized, and filed away for future reference.
Classifying And Predicting
The current study of neural networks based on computer simulation has suggested that they are very good at two specific things: classifying objects and predicting outcomes.
Classifying
It is theorized that the human mind is primarily occupied with classifying things: to take any new sensory input and consider what group of existing memories it "belongs" with, and thereby to recognize it as a member of a group, whose similarities to other items in the group outweigh the dissimilarities.
The author refers to a cat/dog experiment, in which subjects are shown 80 pictures of cats and dogs and are asked to identify which animal is shown. At the beginning of the task, it takes about a second for a person to respond. By the end of the sequence, they have gotten into the rhythm, and can classify them correctly within a tenth of a second. And they do so with 100% accuracy.
(EN: The bit on accuracy, I suspect, comes from the design of the experiment: each image must clearly be a cat or a dog, pictured in such a way that there is no ambiguity. My sense is any ambiguous image would cause the subject to pause. It's still quite remarkable, but I'm wary of the misinterpretation that all recognition is instantaneous and perfectly accurate.)
A significant difference between human cognition and artificial intelligence is that a human being does not need to consciously develop a model or set of rules to identify what is a cat or what is a dog: they simply "know" one from the other without being able to explain the way the tell the difference.
He also mentions experimenting with slipping in a picture of a bear -and even having been instructed to answer "cat" or "dog" the subject invariably calls out "bear" when they see it. This is also a significant difference between human and computer models - the human model is not constrained by an artificial rule-set.
Another variation, done with a bilingual audience, is to ask them to say "cat" or "dog" in one of the two languages they speak, based on an icon in the image. When this second factor is added, subjects still identify the animal correctly, but often fail to use the correct language: the ability to classify an image is a task that is faster and more accurate than the decision of which language to use.
As a follow-up experiment, the author asks the subject if they memorized the eighty images they were just shown - and the subject naturally claims they have not. But then, shown a series of pairs of images, one of which was shown and another of which was not, subjects can clearly recall which ones they had seen, even when the slides moved at a tenth of a second. This demonstrates how quickly memory is formed.
Memory is a form of classification as well - when we see something, even briefly, we file it away as a known object. Within a fraction of a second, we know we have seen it before. Within a few more seconds, we can recall details about the situation in which we had seen it, other things we saw at about he same time, similar things we had seen in other situations, etc.
Predicting
Less is said about the predictive powers of the human mind, and the author provides only anecdotal evidence:
- Upon hearing the first few notes of a song, we know the whole tune
- When a person is speaking slowly, we often can't resist interrupting them because we (think that we) know what they are going to say
- We are amused by jokes and stories that have a "twist" because we already predicted the outcome and are surprised when our prediction is not fulfilled
- Our ability to cross a road is based on the prediction of the future locations of oncoming traffic and our own bodies
- When we sit down to a meal, we can predict according to the sight and scent how the food will taste, and are disappointed if our prediction is not met.
Each of these instances is a form of prediction that is closely related to our power of classification. We draw upon our memory of past experience to decide "what happens next," on the assumption that what will happen in the future will be identical to what has happened in the past.
Marketing Implications
When a customer sees a brand, they classify it based on their past experience. If they have never seen the brand before, they classify it according to similar experiences: they know what milk is, and while they do not know your brand, they identify it as milk and expect it to be similar to their past experience of milk in a general way. If they have seen the brand before, they are familiar with its specific qualities, and expect it to conform perfectly to their past experience.
If they make a correct and positive classification and prediction, all is well - but the danger is that customers will make mistakes. If you sell bleach in a container that is shaped like a milk carton, that has a label that has common elements that the customer associates to milk, then they will think that there is milk inside the container, even if the wording indicates "bleach." And when they discover that it is not milk, their tendency is to blame the marketers for "fooling" them rather than accept that they failed to read the wording on the package.
With that in mind, the task of the marketer is to present a product with consideration of the classification scheme people use. This is the reason that, while there are perhaps a few thousand different brands of wine, they all come in bottles of roughly the same shape. Wine, milk, bleach, and soda all have containers that are expected to be of a specific shape in order for the buyer to recognize the product it contains.
At the same time, the marketer has to find a way to distinguish his brand from similar products - to ensure that the customer will recognize at first sight that there is something different about this particular bottle of wine, it is not like all the others that seem similar, as well as to ensure that the customer who has purchased your brand in the past will be able to easily identify it when they wish to purchase the same brand again.
Similarity leads to adoption by new customers, differentiation leads to repurchase by existing customers. This arises from the processes of classifying (determining what class an unknown item belongs to) and predicting (comparing an item against memory to form expectations).
Any change breaks the association - causes the buyer to fail to recognize your product. One example is that Tropicana experienced a drop in sales of 20% when it changed its packaging, presumably because existing customers no longer recognized it as being the brand they were accustomed to buying. As such, the author cautions brand managers against making packaging changes - it's a fairly obvious way to put their personal stamp on the product and claim credit for future increases in sales, but they would do well to be mindful of the consequences of a brand change.
When designing a new brand, or redesigning an old one, the critical question is not "do you like the design?" but "if you see the design, what do you think the product will be?" The degree to which a person gets this "right" is the degree to which the design will be successful. Where you conform to expectations, your product will be correctly identified by new customers. Where you deviate, your product will be correctly identified by regular ones.
Classification is automatic, and based on past experience - that is, people
"learn" to classify things and use that knowledge to make split-second decisions. Turning again to classifying: if you differentiate yourself significantly from other products, it will be difficult to get new customers to correctly identify your brand as a specific kind of product - but once they have learned the visual cues, it will be difficult for your competitors to convince them that their product is a substitute for yours.
(EN: The soda bottle comes to mind as an example, particularly that of Coca-Cola. IT is similar enough to other soda bottles that a person can identify it as being soda, but different enough that a person will recognize it as a "Coke bottle" and not soda of any other brand.)
This is true of the elements of brand other than product packaging: the logo, images, words, and jingles used in advertising must likewise be considered in terms of their similarity or difference to others. People have an idea of what a television spot for a bank should look like - the general kinds of imagery and themes used - and if a bank deviates too much, its advertising will not register in the memories of the viewers.
Case in point was an advertisement for a bank that used the image of an athlete clearing a high-jump bar, meant to evoke the sense that the bank will help people achieve their financial goals. People who viewed the ad either did not recall it, or recalled the ad but thought it was an advertisement for a popular sports drink.
As an aside, the author mentions trade association advertising - when a given group wants to promote a product without promoting a specific brand. These promotions are successful in increasing sales, but what happens is that the current brand leader benefits disproportionately to other brands.
(EN: I've read that the current "California cheese" commercials have resulted in an increased consumption of cheese in the US market - but the greatest increase in the sale of cheese has been to Kraft Foods, whose main production facilities are not in California, but Illinois.)
Interpretation, Memory Recall, 'Comes To Mind'
Marketers are intently concerned with recall (when a product is mentioned, the subject can name the brand on their own) and recognition (when a brand is mentioned, the subject can relate a memory of having seen it). For commercial purchases, these are the first steps in getting a prospect to know a brand exists, and provide the basis on which the marketer can build interest that will eventually lead to a purchase.
At its most basic level, memory is a survival tool: we are able to take action, or refrain from taking action, because we have memories that enable us to quickly assess whether something we see is beneficial, harmful, or neither. The memories we have must go beyond the mere identification of an object, but to assess whether it is good or bad, and we do so in emotional terms. We do not see a piece of fruit and react by thinking "that mango contains nutrients," but instead we react by feeling "I like mangoes."
This reflects on gestalt: in this instance, the mango gestalt leads to a positive reaction. If a person had eaten one in the past and had a negative reaction (heartburn due to the acidity of the fruit), they might have a negative reaction and avoid it.
The more we encounter something, the more developed our gestalt becomes. A series of positive experiences reinforce one another, and decrease the impact of a negative experience (and vice versa). Aside of desire, our memory becomes more specific and detailed.
It's also noted that as we gain experience, we have more richly detailed gestalt. The analogy is to a cupboard that is overfilled - the greater our experience, the more things there are in the cupboard. If we open the door, things fall out on their own. This is why, in the original example, a person cannot refrain from thinking of an elephant when the word is mentioned - it just falls out of the cupboard when the sound of the word opens the door.