12 May 2009

Linguistic Factoid No. 8: Iconicity

I haven't written any of my linguistic factoid entries recently, the last one being back in February. So I suppose I could write one for today.

And for today, the topic that I have chosen is iconicity.

Now what does this topic have to do with linguistics? The thing is, language is such as marvelous human faculty, and it lets us do many different things. It lets us ask questions, it lets us describe things, and it lets us command other humans. Language is a very interesting tool that us humans have, something that other beings such as animals don't, given the fact that they do not have the ability to do some higher level cognition.

Anyway, what does iconicity have to do with language? Well, iconicity refers to the fact that events are described in the order that they occurred. So, for example, if you ask a schoolgirl what happened in school, she would probably say the following.

I boarded the bus, then I saw my best friend, then we went to class, and the teacher showed us a very pretty book.

The order of the presentation of the events in the narrative roughly corresponds to the actual order of the events. However, that is not the only possible way of stating what happened. Different languages in fact have different ways of encoding events, and sometimes, the language allows us to encode events without following the actual order of occurrence. In English, for example, this language allows us to construct non-iconic discourse by making use of the pluperfect.

John got into an accident. He had been drinking heavily the night before.

In the above case, the drinking event was in fact temporally situated before the accident, but English allows us to construct a discourse where the event order does not match the presentational order of the events.

The curious thing is, not all languages do that. In fact, other scholars argue that there are some languages that only permit an iconic narration of events, one of those languages being Yucatec Mayan. This language happens to be a tenseless one, in the sense that the information as to whether an event occurred in the present or in the past is not encoded grammatically. In this language, narrations always take the order the events happened, and violations of this event order are not permitted.

From a processing point of view, it seems that iconic discourse is easier to understand. Several decades ago, psycholinguists have studied children in how they understand temporal connectives like before or after. It seems that very young children do not comprehend these connectives at all. If I remember correctly, it was Eve Clark who did experiments like these, where she asked very young children to act out certain scenarios, such as the following.

The boy picked up the teddy bear, before/after he patted the rabbit.

If I remember my bibliography right, she found out that very young children use an order-of-mention heuristic when performing the tasks. In other words, whatever action was mentioned first was understood to have happened first, without actually comprehending the meaning of the temporal connective. It is not until ages 4 or 5 that the full import of the temporal connectives are understood.

So there, I suppose one can make up a story why iconic discourse is easier to process than non-iconic discourse. In fact, it is this exact phenomenon that I as a researcher am interested in: what are the processing costs that comprehenders incur whenever they encounter non-iconic discourse? What are the factors that allow people to construct non-iconic discourse? To the extent that non-iconic discourse is possible within a language, how much of a violation is it in quantificational terms? Are there other factors that one can manipulate so that comprehenders would be more likely to expect a non-iconic discourse and therefore reduce the processing cost upon encountering one? If you are interested in this, you can contact me, and I can give you the correct bibliographic sources.

(Marble Garden, from my National Gallery of Art Series)


  1. Wow, never knew about this research. Are tenses used heavily in Indo-European languages? Most Indian languages have 3 variations of each -past,present,future tense.

    Do you know if computers (or robots) have teh capacity to understand such 'complex' instructions? or are they 3 year olds. :)

  2. Final Transit,

    With respect to your question, the answer is indeed yes, Indo-European is known for having complex tense paradigms. In fact, they also have aspect paradigms that combine with the tense paradigms to create more complicated systems. Just to take English as an example:

    I saw.
    I see.
    I will see.
    I have seen.
    I had seen.
    I am seeing.
    I was seeing.
    I will be seeing.

    And I don't think the above is complete.

    I don't really know much about artificial intelligence, although much progress has been done since decades ago.