Introduction

About the experiment

We often look for differences in our surrounding objects to establish a sense of uniqueness that is associated with each of these objects. The experiment is designed to understand how the visual cortex differentiate objects. We, therefore, choose a subset of Telugu characters for this behavioural experiment, wherein, we ask participants to rate the characters based on their perceived dissimilarity between the glyphs that are displayed.

We do not expect participants to read, speak, or write the Telugu language. How such a participant with no knowledge of the language perceives differences is the point behind the experiment. People who know Telugu are also encouraged to participate in the experiment as it will be interesting to compare the perceived dissimilarity of such participants against the participants who have no idea of the language.

Versions of experiment

There are two version of the experiment that were conducted. The only different between the two versions is the scale that was used. The scales are given below:

1. Agreesive scale (α version/ version 1)

   • Very Dissimilar
   • Dissimilar
   • Neutral
   • Similar
   • Very Similar

2. Passive scale (β version/ version 2)

   • Dissimilar
   • Slightly Dissimilar
   • Neutral
   • Slightly Similar
   • Similar

Version 1 was conducted in a monitered fashion at the AI lab, while version 2 was conducted as an online survey.

Instructions

The instructions given for both the version of the experiment are the same. The instructions are given below:

• You will be shown a pair of Telugu characters for a duration of 2 seconds, after which they disappear.
• You are required to rate the pair, in terms of their Dissimilarity.
• Once you have rated the pair, click the next button, to get a new pair
• 30 pairs make up a set. You will have to attempt 10 such sets.
• The estimated time of completion for 10 sets is about 25 minutes.

About the participants

In version 1 of the experiment, all the participants are form university of Hyderabad. We used their registration number as an unique indentifier for these participants.

Since version 2 of the experiment was condcuted as an online survey and, we have used their email IDs as a unique identifier.

Distribution of the Participants

Version α

Data	Value
Participants in $\alpha$ version	23
Telugu participants	7
Telugu readers	7
Telugu speakers	7
Telugu writers	7

Version β

Data	Value
Participants in $\beta$ version	25
Telugu participants	18
Telugu readers	15
Telugu speakers	18
Telugu writers	15