Leslie Linguo

Leslie Linguo is an ESL-oriented chatbot, designed to help students of English as a foreign language to practice their reading and listening skills as well as written communication.

Leslie is able to talk about a wide range of topics which are usually seen throughout the ESL learning process, such as personal information, free time, your town, your home, your friends, films, family, music, work, holidays, the weather, the environment or health. A topic can come up in general conversation, or the student can bring up the topic him or herself.

Grammar and spelling mistakes in writing occasionally disrupt comprehension, but usually they do not. When a student writes "I putted beaker on tripod", for example, he/she will have conveyed the meaning perfectly despite the error.

This same "communication success despite errors" is what Leslie Linguo is about; as long as he understands what the user is TRYING to say, the communicative process will continue. It only breaks down when he does not understand what the student is trying to say, and when this happens he ignores the student's input and attempts to start again by asking a generic, albeit unrelated, question.

Please note that when we refer to "communication" all we are talking about at this level is the exchange of phrases which have meaning within a very short-term context. Real communication is a multiple-way process, and requires that all entities involved understand the concepts, ideas and meanings being presented. This understanding requires a certain level of intelligence, which at present isn’t something that has been adequately demonstrated in a computer program.

Simply put, if you say "HOW ARE YOU?" Leslie knows to answer "I AM FINE, AND YOU?". But Leslie is not fine and he can never be fine. In fact, he does not even know what it is like to be fine. He just knows that in response to the question "HOW ARE YOU?" he must answer "I AM FINE, AND YOU?".

Similarly, if the student mistakenly asks "HOW ARE YOU?" instead of "HOW OLD ARE YOU?" (a relatively common mistake among ESL students) then Leslie will not register the error:

Real life situation:

Student 1: Do you drink alcohol?
Student 2: I am too young to drink alcohol.
Student 1: How are you?
Student 2: I think you mean "how OLD are you".
Student 1: Oh yeah, how old are you?
Student 2: I am 16.

Chatbot situation:

Student: Do you drink alcohol?
Leslie: I am too young to drink alcohol.
Student: How are you?
Leslie: I am fine, and you?

The chatbot doesn't really know what he's talking about; he's basically just spewing out responses to the student's input. This is actually not a bad thing in this case; Leslie's main objective is to keep the communication process alive and if this is possible despite the student's mistakes, whether they be spelling, grammatical or usage, then Leslie will have been a success. In fact, in the best imaginable scenario possible, in many cases a student will realise his or her mistake after reading Leslie's response in the example above and correct it him or herself:

Chatbot situation:

Student: Do you drink alcohol?
Leslie: I am too young to drink alcohol.
Student: How are you?
Leslie: I am fine, and you?
Student: Sorry, I mean how old are you?
Leslie: There is no need to apologise. I am 16.

Peculiar responses and the Turing test
Alan Turing was a brilliant British mathematician. Among his many accomplishments was basic research in computing science. In 1950, in the article Computing Machinery and Intelligence which appeared in the philosophy journal Mind, Alan Turing asked if a machine could think and then went on to ask "If a computer could think, how could we tell?"

Turing's suggestion was, that if the responses from the computer were indistinguishable from that of a human, the computer could be said to be thinking. This field is generally known as natural language processing.

From that idea there have come into existence a number of annual chat bot competitions whose main goal is basically to fool judges into believing that they are talking to a human, and not a robot.

Despite my enormous respect for Alan Turing and many others who have followed in his footsteps, I do not agree with the initial idea that a robot can be considered to be thinking when his responses are indistinguishable from that a human. I believe that a robot can be considered to be thinking when his own botmaster- the person who created and programmed him- is incapable of determining WHY he answers in a certain way.

Therefore, and since we are light years away from that ever happening, I do not believe in competitions where the goal is to fool a judge. In any case, you can't fool someone into believing he or she is actually talking to a real person for very long, and once the veil is lifted most people lose interest. So why bother? Leslie makes it very clear from the very beginning, he's a robot. If you're OK with that, then great! The reason why I write this is because Leslie is not perfect; his responses may sound weird or out of context. Please be patient with him!