The Problem with Blank Slates

Summary:

• Blank Slate (Tabula Rasa) views of human development erroneously claim that we are born devoid of innate mental content.
• From birth, humans already have in place the systems required for rapid learning.
• These systems include those related to object recognition, language, numbers and intentions of others.
• They are then fine-tuned as the infant interacts with the world.

American behaviourist John B. Watson viewed humans as being born tabula rasa, a blank slate, devoid of innate mental content. In one or his most famous quotes, he declared, “Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I’ll guarantee to take any one at random and train him to become any type of specialist I might select – doctor, lawyer, artist, merchant-chief and, yes, even beggar-man and thief, regardless of his talents, penchants, tendencies, abilities, vocations, and race of his ancestors.” He did, however, admit to some over-reaching, concluding that, “I am going beyond my facts and I admit it, but so have the advocates of the contrary and they have been doing it for many thousands of years.” History has, however, proven Watson wrong on many counts, and with the  shift from behaviourism to a more cognitive psychology, along with advances in brain science and computer modelling, it’s become increasingly obvious that the blank slate approach is woefully misleading.  

We often describe this debate as nature versus nurture, although we now have a greater understanding of their interaction with each other (nature and nurture). We are, after all, the result of years of evolution and adaptation, including the development of certain hard-wired systems that interact with the environment and allow learning to happen. This hard-wiring includes the emergence of memory systems that allow us to retain and recall information necessary for survival, or simply hold onto an infinite amount of far less useful information, such as a word by word recitation of our favourite Monty Python sketch. 

But these innate structures also include language systems, structures that allow us to perceive animate and inanimate objects and human faces, and ones that appear to encourage us to differentiate between pro-social and anti-social responses to situations (who we perceive to be good or helpful and bad or harmful). We are even born with the innate ability to understand simple arithmetic. These systems begin to develop even before birth, yet it is within the first months following our emergence into this blooming, buzzing confusion* that this innate, or core, knowledge really begins to take flight.

For example, within only a few months following birth, infants know that some objects only move if acted upon by an external force, while other objects (people and animals) can move without such external assistance. They also understand that objects cannot simply vanish and cannot be in two places at once. 

But how do we know that infants know this?

By observing the behaviour of infants under experimental conditions, researchers have found babies act surprised when objects appear to violate the laws of physics (a phenomenon known, unsurprisingly, as cognitive surprise). By utilising eye-tracking technology, they can determine where babies look and for how long they maintain their gaze, thus measuring the degree of surprise. Many studies have confirmed this, and it is now well understood that, at only a few weeks old, babies already possess deep intuitions of the physical world and show shock when their experiences turn out to be false. 

However, although babies may possess a vast knowledge of the world, they don’t know everything from the start, and it takes a few months for them to understand how the physical world operates. Babies act very much like scientists, testing hypotheses and adjusting their thinking as a result of failures.  

The Good and the Bad

The world is filled with many different objects, and from the first year of life, infants understand that not all objects are the same. Babies don’t show surprise that some objects (people and animals) walk around and move by themselves in ways inanimate objects don’t. By about ten months old, infants are beginning to attribute personalities to people objects. One study conducted by Marine Buon and colleagues from the French National Centre for Scientific Research found that 29 month old toddlers and 10 month old preverbal infants were able to discriminate between the actions of someone whose behaviour was positive (comforting in the study) and a person whose behaviour was negative (harmful). Not only could they tell the difference between positive and negative behaviour, they also displayed a preference to the person who displayed the positive behaviour. It appears they have already formulated some understanding of what it is to be mean or nice, or good or bad, long before they have developed any concrete notions of these concepts or, indeed, have the vocabulary to describe them. 

This might imply that, from a very early stage, infants are able to distinguish between bad behaviour and good behaviour instinctively. But there are still many unanswered questions, of course, that cannot be scientifically and systematically investigated, so any discussion over whether people are born good, bad or indifferent is somewhat superfluous. Indeed, the debate over whether people are born good or bad is best left to those willing to overlook that we are unable to reach such conclusions scientifically.

Numbers

Perhaps the most interesting area of investigation concerns infants knowledge of numbers. We might assume that mathematical skills are learned, but from a very early age, babies appear to instinctively understand number, albeit at a very simple level. For example, if we repeatedly present babies with slides showing two objects, they eventually get bored. Yet, if we add a third object, they will stare at the scene for longer, implying they have detected the change. By manipulating the size, density and nature of objects, we can better investigate infant’s understanding of number. If we show infants several objects and accompany the objects with a sound, say ‘tu’, they show greater interest in the scene if the repetition of the sound matches the number of objects. So if I were to present an infant with four cubed blocks and the sound ‘tu, tu, tu, tu’, they would show greater interest in the scene than if the sound had only two repetitions (see, for example, Izard et al., 2009). Babies, therefore, appear to already possess the ability to recognise and approximate numbers without the need to count.

But can babies calculate? If a 9 month old baby watched me hide an object behind a screen and I then took a second object and placed that behind the screen the baby should assume there are two objects behind the screen – simple enough. However, if I were to remove one of the objects without the baby knowing and then lifted the screen to reveal only one object, the baby would show surprise, one object would be missing (see Koechlin, Dehaene and Mehler, 1997). If I lifted the screen and revealed both the original objects, the baby wouldn’t show any particular interest. Infants only a few months old know the 1 + 1 = 2, and if it doesn’t, they are surprised. The infant knows what is behind the screen, because it’s created an internal mental model of the hidden scene and holds that information in working memory where it can be manipulated by adding or removing objects.

Evidence points to arithmetic as being innate and newborns appear to be able to perceive numbers within a few hours of life (this also true for other animals, including monkeys and ravens). Infants certainly refine these skills over time, but the necessary apparatus is there from birth. 

Both numbers and objects, therefore, represent ‘core knowledge’, those faculties that are present at birth and allow for the development of more complex thought (see, for example, Spelke, 2017).

Language

Nowhere can we see this rapid acquisition of skills more than in language learning. Infants learn their native language incredibly quickly, a feat that would be impossible if each and every word needed to be reinforced, as some behaviourists hypothesised. This is not to say infants are born with a full lexicon or a deep understanding of how to construct sentences conforming to grammatical rules, just that humans are unique in their ability to acquire language (attempts have been made to teach other primates language, including sign language, with little success). 

From birth, babies prefer to listen to their native language compared to a foreign one and there is some evidence that this preference begins in pregnancy. Even babies born two-and-a-half months early respond to spoken language, suggesting this ability begins very early indeed (see, for example, Mahmoudzadeh et al. 2013). 

Babies appear to learn to speak effortlessly and naturally, provided they are raised in an environment where language is spoken. At only a few days old, when babies hear the sound ‘a’ they open their mouths in a way that corresponds with the sound. If they hear the sound ‘e’ they will move their mouths in a different way. Even before they have knowledge of their own mouths, babies are trying to create the sounds they hear. In one interesting study, Stanislas Dehaene and Ghislaine Dehaene-Lambertz scanned the brains of babies while they slept and listed to speech. They discovered that the same brain regions were active in 3 month-old babies as in adult brains when they heard speech in their native language, suggesting that brain organisation doesn’t have to wait for experience to be accumulated in order to process language – the ability already exists.

For us to learn our own language, we need to be able to categorise the sounds that make it up (called phonemes). Newborn babies are able to distinguish between all speech sounds and are actually more sensitive to these sounds than adults. For the first twelve months or so, this sound discrimination is determined by the sounds the infant experiences in the environment, but beyond about twelve months they lose the ability to distinguish between sounds they are not exposed to. 

We can illustrate this process by looking at the Japanese language. Native Japanese speakers cannot distinguish between L and R sounds because the Japanese language doesn’t include them. But Japanese babies can distinguish between L and R, but lose the ability to do this once the sensitive period has passed (so about 12 months). Similar findings have been discovered with other languages (see, for example, Kuhl, 2004).

Without even going into the minutia of how rapidly connections form between our 86 billion neurones, and how each connection represents some kind of learning, it’s clear that the blank slate view is simply untenable. What mechanisms aren’t in place at birth (or even prior to birth) develop rapidly within the first year of life. However, these connections do form as a result of environmental stimuli, so the innate ability to recognise human faces or to understand the motion of objects is fine-tuned as we experience the world. We are wired to learn, and this wiring has been formed during the course of our evolutionary history. 

*I’m using the oft-quoted description from William James’ Principles of Psychology because I like it, but it’s unlikely to be an accurate description of the newborn’s first impressions.

Photo: Scan of a newborn baby’s brain with visualisations of information derived from diffusion-weighted data acquired as part of the Developing Human Connectome Project showing direction of different nerve fibres.