Questions, questions. That’s what’s been vexing me today. I’ve finished the first draft of my literature review – and that feels like such an amazing achievement in itself that gearing up to doing more heavy thinking was no mean feat. But the next thing I need to pin down are the final versions of my research questions. Now, those of you who have done research projects in the past are probably gasping in horror that at this late stage I don’t have them finalised. Well. One thing I’ve learned from this endeavour is that research isn’t as neat and tidy as I once thought it was, or how I now know it should be.
My project is in three phases. The first one started back in 2005 (my god) and was a general exploration of how people felt about near-human agents* and the words that were commonly used to describe them. Very messy, qualitative, very hard to drawn any firm conclusions from the results I got. The second one was a more straightforward experiment testing the effect of turning near-human faces upside down: there’s a well-known effect in psychology where faces are harder to recognise when inverted because this interferes with your overall view of the way the face is arranged. I wanted to see if near-human faces exhibited the same trait or if they were being processed differently, meaning they weren’t as affected by inversion. Then, in the last phase I got clever. Or tried to – I’d come up with a theory that one reason for the uncanny valley effect was that near-human faces were bad at conveying emotional expressions. Either because they weren’t detailed enough or didn’t move accurately, they would set up an expectation of an expression but then not be able to convey it well enough for it to be believable. So I took lots of images of faces where people were intentionally posing happy or sad or frightened or angry expressions and swapped them around, giving the happy faces sad eyes or the angry faces blank eyes. I made lots of different combinations and designed an online experiment where people were presented with the faces and asked how creepy they were, plus whether they could identify the expressions. This was done in a rush when I came back to study after a long break last year and I was just frantically worried about getting the experiment live and getting results in in time to meet my deadlines. It wasn’t well thought through. The first draft of the analysis was also done in a rush and without really knowing what I was looking for. Bad Stephanie, bad PhD student.
But now I’ve been through the literature and have had time to take a deep breath and think, I’m in a better position to be able to pose some sensible questions that this phase can answer.
So, if you’re interested, they are:
- What will the subjective ratings of eeriness be for each of the Facial Expression Blends (FEBs)?
- How will subjective ratings of eeriness for each of the FEBs correlate with participant ratings of their experience of feeling angry, disgusted, frightened, sad, happy or surprised when looking at each image?
- Will the emotions used to create the FEBs be identified accurately, with participants able to identify happiness, neutrality, anger, disgust, fear and sadness when present in the blends?
- Will the identified emotions be the ones displayed in the eyes/top of the face or the surrounding face/bottom of the face?
- Which FEBs will evoke the strongest ratings of anger, disgust, fear, sadness, happiness and surprise?
- Will any of the identified patterns of eeriness, emotion identification and reported experienced emotion be preserved when the FEBs are inverted?
- Will the FEBs be classified in terms of the emotion displayed in the upper part of the face, the lower part of the face, or will they be judged neutral or too difficult classify? The literature suggests that positive emotions are generally perceived from the bottom of the face, negative from the top.
If my brain was fried first thing this morning when I started working on the questions, it’s now curdled, fardled, scrambled and fricasseed. Time, as ever, for a cake. I saw this one mentioned on the Teenbaker blog at the weekend and went for a quick mental walk around the cupboard and fridge – yep, all present and correct. Except when I came to type up this recipe, I realised that in my mental inventory I’d misidentified orange blossom for rosewater. Believe it or not, our tiny Spar-a-like shop at the top of the road had one single solitary bottle of rosewater left on the shelf, looking most out of place next to the pancake mix and own-brand jam. Lucky me.
Pistachio, lime and rosewater cake
Makes one 6″ cake – adapted from this version.
For the cake
- 112g self-raising flour
- Half teaspoon baking powder
- 37g ground almonds
- 50g caster sugar
- 1 eggs
- 25g runny honey
- 125 ml natural yoghurt
- 75 ml sunflower oil
- Grated zest of one lime
- 50g unsalted pistachios, chopped
For the syrup
- 75ml water
- 50g caster sugar
- 1 tablespoon of rosewater
- 50g pistachios
- Juice of half a lime
- Pre-heat the oven to 180C/350F/ gas mark 4. Line the base and sides of the cake tin with greaseproof paper.
- Sift the flour, baking powder and salt into a large bowl. Add the ground almonds and caster sugar and mix.
- Mix the eggs, honey, yoghurt, sunflower oil and lime zest together well in a medium-sized bowl. Make a well in the centre of the dry ingredients and slowly pour in the wet ingredients, bringing them together with a whisk until they are just combined. Add some chopped pistachios to the mixture. Pour this mixture into the prepared tin and bake in the oven for 50 minutes, or until a skewer inserted into the middle comes out clean. Allow to cool in the tin for about 20 minutes.
- Make the syrup while the cake is cooling. In a small saucepan, boil the water and sugar for about five minutes, until it is reduced by half. Add the lime juice and boil for a further two minutes, then cool and add the rosewater to taste.
- With a fine skewer, make holes on top of the warm cake and, with a tablespoon, spoon the syrup all over the top. Scatter the pistachios over the top too, and leave the cake to settle for one hour.
* The term I’ve adopted for that class of things that are almost but not quite human.