Class 8 Intro to Phys Comp – Start of Midterm “Making Meaning: By the Mouthful”


This project is a game that lets people create new words to fit meanings that don’t have words in the English language… yet.

To start playing, people take apart a human mouth, and use each part to pick out consonants and vowels to create random words. There’s more to the game described below, but to start here is my prototype of this first aspect of the game.


Larger Picture

I decided to work on this idea because I DIDN’T want to use a special sensor, so that I could focus on the interaction design itself.

As for the interest in words, I’ve also always liked the game Fictionary because you can create fake definitions for obscure words in the dictionary. And who doesn’t enjoy learning about other language’s words that don’t exit in English? Some new words I can relate to, but others describe feelings I haven’t even had before.

To be honest, I also wanted to make something soft that you could press or squish that would change something on screen.   Hence the stuffed felt!

You can see that ultimately I want the game pieces, which form a mouth, to not only fit together but also come apart so that you can press them to choose different consonants and vowels.

I’m hoping taking apart the mouth, and how soft and pliable the parts are, gives hints as to how to get different kinds of vowels and consonants. For example, you can create “nah” by pressing the tongue and mouth pieces, because those are the parts of the mouth most required to create that sound. But for vowels such as “ooh” and “uhh”, you only need to press the mouth.  I need to do some user feedback testing to see how much this gets across without explanation.


Below is the whole game as a user map. People can decide where they’d like to start. They can either:

  • Create a new word, record and spell it.
  • Define the meaning of a new word.
  • Vote on the top words and view the best ranked words.
  • Request a new word for a meaning they’ve always wished there was a word for.




I started by creating a paper model of the mouth.  Below is the mouth, teeth and tongue, all fit together. Ultimately, magnets might work to keep it all together. Or the friction of felt might be enough on its own.


Here they are as game pieces.


Then I played with soft circuits to create felt versions of these paper game pieces. Thank you to Hannah/Pulsea on Instructables for their directions on soft circuits, such as this one, and their collective Kobakant’s  very extensive library of soft things. They went to MIT Media Lab and were part of the High Low Tech research group! Very cool.

Here I am testing a tongue with a coin battery circuit.

giphy_squeezing tongue

As for how it works, the tongue has three layers of felt. In between two layers are the circuit parts. In between the other two layers you’ll find only fiber fill.

The circuit is made of a a sandwich of Velostat between two pieces of conductive fabric. Velostat is a kind of paper foil with carbon in it, making it conductive as well. The conductive fabric I found had adhesive on the back, which saved me the extra step of ironing on interfacing fabric to adhere it to the felt.

I’m assuming the circuit is completed when the “tongue” button is pressed hard enough to send electricity from one piece of conductive fabric through the velostat to the other fabric. Check out Pulsea’s instructables to understand how to shape the conductive fabric so they don’t accidentally touch, which creates a short circuit.

Half the inside of the tongue, showing the circuit’s conductive fabric on the bottom. I used alligator clips. I need to look into XBee Radios or anything else that would make these mouth parts wireless. 


I pinned paper stencils to the felt to make it easier to cut with fabric scissors.


As for the code, I used serial communication to connect the value readings of my analog sensors to my p5.js sketch.

Here is my arduino code, which uses a handshake method to slow down the transfer of data to only when the p5.js sketch asks for it and reads and prints serially the analog values of each sensor.

And here is my p5.js code. This is only my sketch.js file. For my final iteration, I’ll learn to post my code on github as a complete folder with all files.

I’m wondering if there is a better way to store my logic for which sensor combinations play which sounds. Ultimately it could be 24 consonants and 20 vowel sounds in combination with each other!?!

I also need help improving the code to allow a user to press two sensors and get only one result, instead of triggering each individual sensor’s results. Right now, it sort of works but sometimes you end up hearing as many as three sounds.

Finally, here is my circuit on Fritzing. The red line is Fritzing’s interpretation of a soft circuit…

Making Meaning_01_bb