Yerkish

Yerkish

Primary Disciplinary Field(s): Cognitive Science, Primatology, Comparative Psychology, Linguistics

1. Core Definition

Yerkish is a highly structured, symbol-based communication system designed primarily for non-human primates, specifically chimpanzees, to facilitate linguistic interaction with human researchers. Developed in the early 1970s at the Yerkes National Primate Research Center in Atlanta, Georgia, Yerkish utilizes a set of visual symbols known as lexigrams. These lexigrams are arbitrary in nature; that is, they are not necessarily iconic depictions of the objects or concepts they represent, but function as abstract equivalents of words. The system was meticulously constructed by psychologist Ernst von Glasersfeld, who focused on the structural and functional requirements necessary for a successful artificial language system (ALIS) capable of being acquired and manipulated by primates. The design philosophy emphasized syntactic order and semantic clarity, aiming to test the cognitive capacities of chimpanzees for language acquisition beyond simple conditioning.

The communication environment for Yerkish typically involves a computer-controlled keyboard where the lexigrams are displayed and accessed. When a subject—such as a chimpanzee—presses a sequence of keys, the corresponding lexigrams illuminate, and often, an auditory translation or synthesized voice output is generated. This immediate feedback mechanism reinforces the association between the symbol, the concept, and the resulting action or consequence. Yerkish is significant because it allows researchers to rigorously track the sequences and grammatical structures used by the primate subject, providing quantifiable data on sentence formation and concept usage. The foundational work utilizing Yerkish was conducted under the leadership of comparative psychologist Duane Rumbaugh, particularly through the ambitious language project involving the chimpanzee named Lana.

Unlike earlier attempts at teaching sign language to primates (e.g., Project Washoe), Yerkish provided a fixed, non-ambiguous input and output mechanism, eliminating reliance on human interpretation of manual signs, which could be subjective. The system required the primate subjects to understand and produce specific, ordered sequences of symbols to achieve desired outcomes, thus testing their understanding of rudimentary syntax. The success of Yerkish demonstrated that chimpanzees were capable of linking abstract symbols to meanings and using these symbols productively and creatively, challenging traditional behaviorist views regarding non-human animal cognition and communication potential.

2. Etymology and Historical Development: The Lana Project

The term Yerkish is an eponym, derived directly from the institution where the system was conceived and first implemented: the Yerkes National Primate Research Center. The project began in 1971 with the central goal of investigating whether chimpanzees could acquire and manipulate language in a structured, controlled environment. Duane Rumbaugh, along with his colleagues, initiated the project, commissioning Ernst von Glasersfeld, a linguist and philosopher known for his work in constructivism, to design the artificial language itself. Von Glasersfeld’s design focused on creating a language where the rules of construction (syntax) were explicit and unambiguous, ensuring that any perceived linguistic behavior was a function of the chimp’s understanding of the system, rather than simple rote memorization.

The initial and most famous application of Yerkish was the Lana Project. Lana, a female chimpanzee, became the primary subject in this study, which utilized a novel computer interface called the Language Analogue (LAN) system. The LAN system integrated the Yerkish lexigram keyboard with a computer, allowing for precise recording of Lana’s communicative output, including sequence, timing, and errors. The development was critical because it moved primate language research away from purely observational methodologies into a verifiable, computer-mediated experimental framework. Early stages involved teaching Lana to request objects and activities by pressing corresponding lexigrams in a specific order, such as “Please machine give piece of apple.”

As the project progressed, Lana successfully mastered complex tasks, including distinguishing between different types of food, identifying researchers, and even asking questions. A pivotal development occurred when Lana began to spontaneously construct novel sentences by combining known lexigrams in grammatically correct ways, demonstrating productivity. Furthermore, Lana showed the ability to correct her own errors and to use a specific lexigram sequence—”question mark”—to request the name of an unfamiliar object, a remarkable indication of metalinguistic awareness and the capacity to utilize abstract tools for knowledge acquisition. This historical achievement established Yerkish as a benchmark for subsequent studies in comparative psychology, particularly those involving symbol-based communication systems.

3. The Lexigram System

The core component of Yerkish is the lexigram. Lexigrams are two-dimensional geometric figures that serve as the fundamental units of meaning (morphemes or words) within the system. Designed by Von Glasersfeld, these symbols were intentionally abstract; unlike pictograms, they do not visually resemble the objects they denote. For example, the lexigram for “apple” bears no visual resemblance to the fruit, requiring the subject to form a purely arbitrary association between the visual pattern and the semantic concept. This abstraction was deliberate, designed to test the true symbolic capacity of the primates, ensuring they were not simply reacting to visual cues but were engaging in genuine symbolic representation.

The full vocabulary of the Yerkish system consisted of numerous lexigrams, each representing nouns (objects, people), verbs (actions), adjectives, and grammatical particles. These lexigrams were arrayed on a dedicated keyboard, often referred to as the “Yerkish keyboard” or the “LAN board.” The typical layout involved 100 or more keys, each associated with a unique symbol. When pressed, the symbols would appear in a linear sequence on an adjacent display screen, forming a “sentence.” The physical arrangement and color-coding of the lexigrams often served pragmatic functions, helping the subjects locate and differentiate the symbols, but the meaning itself was entirely based on learned convention.

Crucially, Yerkish employed a strict syntactic structure, based loosely on English grammar, requiring subjects to press lexigrams in a specific order to constitute a meaningful request or statement. This rule-bound nature was necessary to prevent ambiguity and to test the primates’ ability to adhere to grammatical rules—a defining feature of human language. For instance, a typical request might follow the structure: Request (e.g., Please) + Subject (e.g., Machine) + Action (e.g., Give) + Object (e.g., Water). Deviations from this sequence were registered by the computer as errors, ensuring that the acquired language was functionally syntactic, rather than merely a chain of unrelated requests.

4. Methodology and Implementation

The implementation of Yerkish relied heavily on a sophisticated computer system, making it one of the first successful applications of technology in comparative psychology research. The methodology focused on intensive training using operant conditioning principles, but situated within a communicative context. Subjects were initially taught the functional use of individual lexigrams, primarily through associating the symbol with a tangible reward or action. For instance, pressing the lexigram for “banana” would immediately result in the dispensing of a banana. As competence grew, the researchers shifted the emphasis from simple association to understanding the functional relationships between symbols.

A key methodological technique was the use of complex discrimination tasks. Subjects had to demonstrate not only that they recognized a symbol, but that they understood its role in a sentence structure. For example, the subject might be required to differentiate between “Tim give Lana water” and “Lana give Tim water,” proving their comprehension of subject-object relations defined purely by the sequence of the lexigrams. The computer system served as an impartial, consistent judge, providing immediate feedback and detailed logging of all interactions, which allowed researchers like Rumbaugh to perform robust statistical analysis on the linguistic output.

The environment was structured to encourage active communication. The primates’ access to resources (food, toys, interaction with humans) was frequently mediated through the Yerkish system. If a chimpanzee wanted a specific item, they had to successfully compose the correct lexigram sequence to request it. This created a strong, intrinsic motivation for the subjects to master the system, viewing the lexigrams not just as stimuli for rewards, but as tools for controlling their environment and communicating intent. This communicative drive, combined with the rigorous structure of the language, formed the basis for the advanced capabilities later demonstrated by subjects like Lana and, subsequently, other primates studied at the Yerkes center.

5. Key Findings and Successes

The primary success of the Yerkish system was its demonstrable proof that chimpanzees could acquire and use an abstract symbolic system syntactically. The most compelling evidence came from the abilities demonstrated by Lana. She acquired a vocabulary of over 100 lexigrams and consistently formed grammatically correct sentences to request items, obtain human attention, and engage in social routines. Her capacity for semantic mapping—the ability to correctly associate abstract symbols with concepts—was clearly established.

Furthermore, Lana demonstrated displacement, a critical feature of human language. She could talk about objects that were not immediately present or request actions that would occur in the future, proving that her communication was not merely tied to the immediate physical environment. More impressively, Lana exhibited productive language use. When presented with an unfamiliar carbonated beverage (soda), she reportedly referred to it as “Coke that black good,” demonstrating the ability to combine known descriptors (lexigrams for “Coke,” “black,” and “good”) into a novel, yet descriptive, phrase, indicating a flexible understanding of the language’s semantic components.

The Yerkish method was also successfully applied to other primates, most notably the chimpanzee Sherman and the bonobo Austin, also under Rumbaugh’s direction. Sherman and Austin were trained not just on single lexigram requests, but on the concept of classification and naming. They learned to label items based on categories (“food,” “tool”) and demonstrated the ability to communicate information to each other using the Yerkish keyboard, such as informing their partner which tool was needed to obtain a specific reward. This successful inter-subject communication using Yerkish strongly supported the view that the system represented a viable channel for symbolic thought transfer among primates.

6. Theoretical Underpinnings

The philosophical foundation of Yerkish is deeply rooted in the work of its designer, Ernst von Glasersfeld, particularly his concept of Radical Constructivism. Von Glasersfeld argued that knowledge is not passively received but actively constructed by the cognizing organism. In the context of Yerkish, this meant that the chimpanzee subject was not merely mimicking human instruction, but was actively building a semantic and syntactic framework based on the functional utility and consistent consequences of their lexigram choices. The language system itself was designed to be a tool for the construction of meaning.

This constructivist approach contrasts sharply with purely behaviorist interpretations, which would suggest that Lana was simply responding to stimuli chains (S-R). Von Glasersfeld and Rumbaugh argued that the chimpanzees’ novel constructions and spontaneous errors corrections indicated an underlying cognitive model—a representation of the language rules—that the primate was utilizing to generate meaning. The systematic, computer-controlled nature of Yerkish provided the necessary consistent environment for the chimpanzee to develop and test hypotheses about the language structure, thereby constructing their own understanding of its rules.

Furthermore, Yerkish provided strong evidence in the ongoing nature-nurture debate concerning language acquisition. While Noam Chomsky posited that human language relies on an innate, species-specific linguistic structure (Universal Grammar), the achievements of the Yerkish-trained primates demonstrated that complex symbolic and syntactic capabilities could be acquired through structured learning and environmental interaction. While Yerkish did not prove that chimpanzees possess innate human language capacity, it undeniably expanded the understanding of non-human linguistic competence, suggesting that the building blocks of syntax and symbolism might exist across species, capable of being unlocked through appropriate training environments.

7. Criticisms and Limitations

Despite its methodological rigor, Yerkish, along with all ape language projects, has faced significant criticism from the linguistic community. The most prominent critique, often leveled by proponents of the Chomskyan school of thought, is that the output generated by Yerkish subjects does not constitute true language, but rather highly sophisticated forms of operant conditioning. Critics argue that the primates learn to associate specific symbol sequences with desired rewards, effectively treating the lexigram sequence as a complex tool-use behavior rather than a generative linguistic system. The communication, in this view, is primarily pragmatic and petitionary (requesting), rather than declarative or expressive.

A second major limitation concerns the nature of the syntax observed. While Lana could adhere to the rigid order required by the Yerkish computer, critics question whether she truly understood the abstract grammatical relations (e.g., subject, verb, object) or merely the necessary sequence required to trigger the machine’s reward function. The syntax mastered by the chimpanzees was highly constrained and lacked the recursive complexity and infinite generative capacity characteristic of human language. Furthermore, researchers had to be extremely careful to avoid the Clever Hans effect, where subtle, unconscious cues from the human experimenters might guide the primate’s choices, although the automated computer interface of Yerkish greatly mitigated this risk compared to sign language studies.

Finally, the ecological validity of Yerkish is often debated. Since Yerkish is an artificial, human-designed system implemented in a laboratory setting, it does not reflect the natural communication methods of chimpanzees. While the system effectively tests certain cognitive boundaries, the practical significance of teaching a primate a non-natural language remains contested. Nevertheless, subsequent research, particularly with the bonobo Kanzi, who acquired a similar lexigram system observationally (without explicit reward training), provided stronger evidence that symbol acquisition among primates can transcend simple conditioning, lending greater weight to the cognitive relevance of the Yerkish approach.

Further Reading

• Yerkish and Lexigrams (Wikipedia)
• Rumbaugh, D. M., Gill, T. V., & von Glasersfeld, E. C. (1973). Reading and sentence comprehension by a chimpanzee (Pan troglodytes). Science.
• Lana (chimpanzee) and the Language Analogue System
• Von Glasersfeld, E. C. (1978). The Conception of Language.