www.e ol ingci ies.o g
Limna and he Shi ing Epis emologies o A -Ma ke Value
D . Leande Gussmann, A x Science School o T ans o ma ion (JKU, Linz, Aus ia).
Con ex and Resea ch Ques ion
The Limna app o sma phones, ma ke ed as a “pocke a
ad iso ,” ma ks a shi om connoisseu ial, expe -d i en
alua ion owa d da a-cen ic alida ion. T adi ional p icing
es ed on opaque, socially nego ia ed “p i a e codes”. Limna
ins ead d aws on A ac s’ knowledge g aph o oughly
800,000 a is s and mo e han 1 million exhibi ions o p oduce
p ima y-ma ke p ice bands. This esea ch asks: How does
his hyb id epis emology, which claims objec i i y while being
ained on en enched hie a chies, ede ine alue in oday’s
a ma ke ? I examine whe he algo i hmic alua ions
ans o m epis emic au ho i y o epackage exis ing powe
and p es ige sys ems.
I Say: A en ion as Cu ency
Tensions: Black Box Vs. Reali y
Algo i hmic alua ion ools demand c i ical da a li e acy. Use s mus
in e oga e p oxy a iables (e.g., exhibi ion coun s as quali y p oxies)
and ecognise ha ankings de i e om cu a ed social ne wo ks. These
sys ems a e pe o ma i e—p ice bands can shape ma ke expec a ions
and ou comes. Fu u e wo k should mo e beyond in e ace analysis o
e hnog aphic ieldwo k: in e iews wi h de elope s, collec o s and
galle ies could e eal whe he hese “p obabili y engines” empowe
new buye s o en ench “a en ion eli es.” Compa a i e s udies ac oss
pla o ms and mic o-cases o indi idual a is s would illumina e how
algo i hmic ools edis ibu e epis emic au ho i y in a ma ke s.
Me hod: The “Sociological Scanne ”
I econs uc Limna’s da apoin pipeline h ough public
documen a ion and in e ace analysis. A ac s manually
cu a es and e i ies exhibi ion da a be o e assigning “a en ion
poin s”. Limna hen inpu s hese sco es, plus a wo k
dimensions and sales his o y, in o a machine-lea ning model.
This p ocess is a o m o commensu a ion, u ning
he e ogeneous e en s in o compa able numbe s. Empi ical
wo k shows ha social me ada a, comp ised o ins i u ional
ies and p io sales, explains abou 73 % o p ice a iance,
while isual ea u es accoun o only 5.5 %. Acco dingly,
Limna beha es less like an assesso o in insic quali y and
mo e like a scanne o a en ion capi al.
In e ace App aisal
Figu e 5: O iginal image a 100%, enla ged 200% and 400%.
They Say: The P omise o Accu acy
Limna’s ma ke ing p omises anspa ency and educed isk
o newcome s. I s FAQ s a es ha unde 20 % o p ice
es ima es di e ge om galle y quo es. P omo ional ma e ial
ci es co e age o hund eds o housands o a is s, o e a
million exhibi ions, and six een housand galle ies. Founde s
ame he app as cla i ying he “basic mechanisms” o a
p icing and democ a ising ma ke knowledge, implying ha
isualised ca ee ajec o ies can challenge he opaci y o
p i a e p icing.
In e ace Disco e
Con ac In o ma ion
Leande Gussmann
A x Science – School o
T ans o ma ion
JKU Linz
Aus ia
Email:
leande [email protected]
A gap exis s be ween pla o m he o ic and pe o mance.
Du ing A Basel 2021, jou nalis s epo ed ha Limna alued
a Conny Maie pain ing a €4 500, whe eas he galle y p ice
was mo e han h ee imes ha . Such e o s p obably happen
less equen ly now. Such disc epancies, ne e heless,
highligh ha models ained on his o ical da a ha e limi s;
hey canno cap u e hype o sudden ma ke shi s. The
sys em’s emphasis on es ablished ins i u ions isks
ep oducing geog aphical and gende biases baked in o
his o ical da a. E en hough Limna ca ego ises a is s (e.g.,
" esh", “eme ging,” “es ablished”), i s unde lying hie a chy
s ill a ou s canonical enues.
Da a Li e acy & Fu u e Resea ch
Call o ac ion
D awing on F anck’s economy o a en ion, I con end ha
Limna quan i ies cu a o ial in es men . Museums, biennials
and galle ies lend epu a ion; hei exhibi ions gene a e da a
poin s ha eed an “a is ac o ” and p oduce p ice bands.
The app hus unc ions as a judgmen de ice (Ka pik, 2010)
and pa icipa es in me ic go e nance: a he han disco e ing
in insic alue, i no malises speci ic o ms o isibili y. A is s
a e incen i ised o maximise measu able ins i u ional
ecogni ion— h ough momen um and cul u al ecogni ion
sco es—a he expense o less quan i iable expe imen a ion,
he eby eo ien ing ca ee s owa d anking and ne wo k
cen ali y.
T ea pla o ms like Limna no as neu al p ice engines bu as
pe o ma i e judgmen de ices: ques ion hei p oxies, ask
who and wha hei da ase s lea e ou , and insis on
anspa en me ics and accoun able use whene e algo i hmic
alua ions in o m collec ing, cu a ing, o policy.
