Ned Block - Why Do Mirrors Reverse Right/Left But Not Up/Down

Journal of Philosophy, Inc.
Why do Mirrors Reverse Right/Left but not Up/Down Author(s): N. J. Block Reviewed work(s): Source: The Journal of Philosophy, Vol. 71, No. 9 (May 16, 1974), pp. 259-277 Published by: Journal of Philosophy, Inc. Stable URL: http://www.jstor.org/stable/2024963 . Accessed: 16/12/2012 17:45
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Journal of Philosophy, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Philosophy.
http://www.jstor.org
This content downloaded on Sun, 16 Dec 2012 17:45:55 PM All use subject to JSTOR Terms and Conditions
THE JOURNALOF PHILOSOPHY

VOLUME LXXI, NO. 9, MAY .-
i6, I974
.~~~~~~4.
it is that theyare asking. I shall argue that there are many interpretations the question, and that they fall into two of generalcategories:(1) interpretations which the correct on response to "Why do mirrors... ?" is: they don't; (2) interpretations on whichmirrors reverse do and not up/down,but whereone right/left can explain why theydo throughan examination of the concepts 'right','left', 'up', and 'down'. of One example of an interpretation type 1 is the following.Suppose I have a placard such as that picturedin A.
Tq33J\TH3Ih5 aHMWIAV3 Oa 25OA5105IIM YHW

*
WWOU\qU TOVI TUff
knowjust whatquestion EOPLE who ask thisquestion rarely
(A)
WHY MIRRORS REVERSE RIGHT/LEFT DOBUTNOT UP/DOWN?
Why is it that the mirrorimage of this placard, viz., B, reads from
(B)
(B
oa 2lOHflIM YHW 3fl3V3R TI31\THDIfl NWO\q9U TUO TON
rightto left instead of left to right,but still reads up to down as before?I shall call this interpretation "Why do mirrors.. . ?" of interpretation for "reading." When contrasting 'R' the notion of reversal deployed by interpretation with other notions of reR versal,I shall call it "reversalr." (All the notionsof reversaland the interpretations our original question discussed in this paper of are listed below in the appendix.) What is puzzlinghere is that the
*I am indebtedto GeorgeBoolos, Michael Devitt,Hartry Field, Paul Horwich,Jerry Katz,and RichardMillerforcomments an earlierversion this on of paper. 259
260
THE JOURNAL OF PHILOSOPHY
mirror seems to treat the horizontal axis differently from the verticalaxis. But notice that B is by no means "the" mirrorimage of A. For example,anotherpossible mirror image of A is C.
(C)
The reason B appears to most of us to be the mirrorimage of A is as follows.Imagine you are facing the mirrorwhile holding the placard in frontof your face reading it in the normal way. You then rotate the placard 1800 so that it faces the mirror.Now most of us tend to turnthe placard to face the mirror rotatingit about by itsvertical(y) axis. Thus we see B in the mirror. But if we rotatedit about its horizontal (x) axis, we should see C. But C reads normally the x axis (leftto right)and abnormally on on the y axis (down to up). So if this way of turning the card were more natural than the other way, people would perhaps be temptedto ask: Why do mirrors reverse up/downbut not right/left? Suppose you are in a room,one wall of whichis a mirror; attached to the opposite wall is the placard A. You stand facingthe placard. You then turn to the mirror.What you see is B. But the puzzle now seemsreborn,since the placard and the mirrorare stationary, being attached to the walls of the room: neitherhas been rotated. But notice that though the placards have remained stationary, you have turned about your vertical axis (see Figure 1). Had you
MHA WIHBO2 2E ICHH\rEL DO IHEAEBE
Doli nb\DOMNS mO1
FIGURE
1. Viewof theroomfrom side. thie
WHY DO MIRRORS REVERSE RIGHT/LEFT
BUT NOT UP/DOWN?
26I
\ \Zf\\SSORS REVERSE /
\
\
/\6AtAFTp 8UT/OT MR~\/DOW?
n~~i6\o0^flKV\6
FIGURE
2. View of the room fromthe top. You see the reflection just as the person standingon his head does.
turned by flipping over backward or by bending down and looking throughyour legs, you would have seen C (Figure 2). ways of describingthis situaNow there are two quite different tion. One way would be to say that the mirrorreversesthe placard right/left not up/down in one circumstance-ifyou rotate the but placard (or yourself) 1800 about the y axis (case B)-whereas in the mirror reversesthe placard up/down and not right/left another circumstance-if you rotate the placard about the x axis (case C). A second descriptionis: the mirrorappears to reverse right/left one case, whereasit appears to reverseup/down in the in nor up/ other,but actually the mirrorreversesneitherright/left down in eithercase. I incline toward the latterdescriptionfor two reasons.First,to say that the mirrorreversesthe placard right/left is to say that the mirrorimage of the placard reads right to lefteven if one specifies that the placard be rotateda certainway. And to say the mirrorimage of the placard reads rightto leftis to presuppose (falsely) that there is a unique mirror image of the placard.1 Secondly, to say that the mirroractually reversesright/
1 It might be objected that (1) the phrase 'the mirror image of the placard when it is rotated about the y-axis' does (uniquely) refer,and thus that it is true that (2) the mirror reversesthe placard right/left when the placard is rotated about the y axis. Now (1) is true, but (2) is ambiguous; and the interpretation of (2) that (1) supports is not the interpretationthat provides an answer to the original question. The ambiguityin (2) is an ambiguityin the scope of 'reverses'. One way of interpreting is as assertingthat the mirrorreversesthe y-rotated (2) (2) placard. On this interpretation, is true but irrelevant: for the original questioner wants to know why the mirrorreverses the placard, not why the mirror (2) The other way of interpreting is as reversesthe y-rotatedplacard (right/left).
262
a left but not up/down (even if one specifies y-axisrotation) is to implythat the mirrortreatsthe up/down axis of the placard differaxis. But the mirrordoes not treat the ently from the right/left It two axes differently. is we who do so by choosing the up/down are axis as the axis of rotation.If thesearguments right,the correct R responseto interpretation of "Why do mirrors... ?" is: theydon't.
II
reAnother (closely related) interpretationon which right/left versal is illusorywas expounded by D. F. Pears in his "Incongruity 2 of Counterparts." Pears asks us to "imagine that I look into a accurately portraitof myself mirrorand then produce a full-length image.Next suppose thatI go round behind painted over the mirror the mirrorand face the back of it. Suppose also that the mirroris made of some flexibleplastic material.Then I can put the portrait of my face on my face like a mask, and the portraitof the rest of my body on the rest of my body like a complete suit of clothes" (78). Let us call the processof puttingthe plastic on my body "getting into myportrait."Why is it that,when I get into my portrait, (The right/left? but fails to fitvis-'a-vis top/bottom, it fitsvis-A-vis puzzle assumes that I do not have perfect bilateral symmetry; imagine for example, that my left arm has a cast on it.) G" of I shall call this interpretation the puzzle "interpretation this into one's portrait")and (whencontrasting with ('G' for"getting othernotionsof reversal)I shall call the notionof reversalit deploys Pears' solution consistsin pointingout that the appear"reversalg." here owes to the fact that the normal way of ance of asymmetry gettinginto the portraitinvolvespivotingon one's verticalaxis in order to face the back of the mirror."But suppose that instead in of turning the verticalaxis I turnedin the horizontalaxis about which we pivot when we turnhead over heels: suppose that I stood It on myhead behind the mirror. is equally clear that,if I got into my portraitin thisway,leftand rightwould not be reversedwhile top and bottomwould be reversed"(loc. cit.).3Of courseif I got into the portrait this way it wouldn't fit very well, since I would be puttingthe "legs" on myhead and vice versa.But thisbad fitoccurs
assertingthat the mirrorreversesthe placard right/left-in a given circumstance (y rotation). But this interpretationpresupposes falsely that 'the mirror image of the placard' refers. I'hus on one interpretation(2) is not relevant, and on the other interpretation is not true. it 2 Mind, LXI, 241 (January1952): 78-81. 3 I would say "appear to be reversed"where Pears says "reversed"; see the last paragraph of this section.
WHY DO MIRRORS REVERSE RIGHT/LEFT BUT NOT UP/DOWN?
263
merelybecause I am not symmetrical above and below the waist. So one correct response-though one Pears himself apparently would not accept (see the paragraph afternext)-to Pears' version of "Why do mirrors . . ?" is: theydon't, and the appearance that . theydo owes to the fact that in "gettinginto one's portrait"one would normallyturn around one's verticalaxis. Further,the possibility of gettinginto one's portrait upside down is obscured by the fact that we are not symmetrical about the plane that cuts a standingfigure at horizontally the waist; werewe symmetrical about thisplane and were it natural to get behind thingsby flippingupside down,people mightwonderwhymirrors reversed up/down but not right/left. Bernard Mayo 4 puts Pears' point slightlymore generally.An asymmetrical object and its reflection cannot be made to coincide one is superimposedon the other.(Such a pair of objects are when said to be "enantiomorphs.")But which axis bears the failure of coincidence (and thus appears to be reversed)depends entirelyon how one rotatesthe object and its reflection beforeone attempts the superimposition.Any axis one chooses can be made to appear reversed means of a suitablerotation.In terms Pears' example: by of if one naturallychose to get into one's portraitby walking behind it and then backing into it (i.e., no rotationat all) it would be the front/back axis that appeared to be reversed, not the right/left or up/down axis.5 And if one backed into it standingon one's head, all threeaxes would appear reversed. as With reversalg withreversalr, thereare two quite different ways of conceptualizingthe situation. One way, apparentlyfavored by reverseright/left Pears, is to say that mirrors (and not up/down6) ifone imaginesgettinginto one's portrait the usual way,but that in mirrorsreverse up/down (and not right/left)if one imagines gettinginto one's portraitfrontward standingon one's head. The alternative conceptualization says that mirrors reverse neither nor up/down in either case, even though mirrorsmay right/left appear to do one or the other,depending on which operationsare imagined to be involved in gettinginto one's portrait.
4 "The Incongruity Counterparts," of Philosophy Science,xxv, 2 (April of 1958):109-115. 5 Superimposition withoutrotationproducesfailure to "fit" only on the frontward/backward (whenfacingthe mirror). axis This maylead some people to suppose that the mirrorreallyreverses only frontward/backward. to But supposethiswouldbe to introduce newsenseof'reverse'-asenseuninteresting a in thatit givesrise to no puzzlein the first place. See footnote 15. wouldactually moreappropriate be 6'Top/bottom' herethan'up/down'.
264
As with the firsttype of reversal,I incline toward the second right/ typeof description.First,to say the mirroractually reverses left is to say that gettinginto one's portraitrequires that the porAnd to say this is to presuppose right/left. trait fail to fitvis-'a-vis i.e., that thereis one falselythat 'gettinginto one's portrait'refers, and only one typeof gettinginto one's portrait.(See the analogous discussionat the end of part i, especiallyfootnote1.) Secondly,to but reverses right/left not up/down is to imply say that the mirror from one's that the mirrortreatsone's up/down axis differently axis-even if one specifiesthat one is imagininggetting right/left into one's portrait by turning on one's up/down axis. But the It mirrordoes not treatthe two axes differently. is we who do so by pickingone axis as the axis of rotation. as As interesting the two versionsof "Why do mirrors... ?" so far discussedare, I doubt thatverymany people who have worried ... about why mirrors have had either of them in mind. I claim: (a) Most of thosewho have any views at all about frontward/backward reversalfeel, at least initially,that mirrorsdo reversefront(b) (whenwe face themirror). Most people who have ward/backward worriedabout the problemprobablywould not feel (even initially) if frontward/backward theyunderor reverser reverseg that mirrors are. and reversalg stood what reversalr In short,although mirrorsdo appear to reverse,they probably frontward/backor would not appear eitherto reverser to reverseg people who have worriedabout reversalprobablyhave not ward. So My or (in so doing) worriedabout eitherreversalr reversalg. evidence an informalsurvey.This type of evidence is not for (a) is simply easily obtainable for(b), since most people who have worriedabout interpreamong the different whymirrors.. . have not distinguished argumentsI have already given that tations of the question. The can easily right/left nor reverseg mirrorsactually neitherreverser frontnor neitherreverser reverseg be adapted to show that mirrors point is not about what mirrors ward/backward.But the present actuallydo, but ratherabout what theyappear to do. In the restof do this section,I shall back up (b) by arguing that mirrors not apeven in the way frontward/backward or pear to reverser reverseg right/left. and reverseg to theydo appear (misleadingly) reverser but right/left Reversal,. The reason mirrorsappear to reverseg an obnot up/down is that it is natural to attemptto superimpose by ject on its reflection turningthe object (perhaps only in one's imagination) about its y axis. As Mayo points out, in order for it it frontward/backward,would have to reverseg to seem thatmirrors
BUT NOT UP/DOWN?
265
be natural to accomplish such a superimpositionwithout such a rotation(and withouta combinationof rotationsto the same effect). In Pears' terms,it would have to be natural to get into one's portraitby backing into it. But what would have to be natural is frontward/backdo not natural-so mirrors not appear to reverseg ward. clear what would even count as frontReversal,.It is not entirely ward/backwardreversalr(real or apparent). Imagine (if you can) one pointed that,when one wanted to read a placard in the mirror, the back of it toward the mirror-so that what one saw was the blank back. Perhaps in this situation we would have apparent reversal. More intelligibly: imagine that frontward/backward rectangularsolids that made use printed objects were transparent of all threedimensions.The conventionfor directionof reading a solid mightbe: top frontrow (left to right) then simple four-row top back row, then bottomfrontrow, then bottomback row. Now if one were to read such a printedsolid in the mirrorby pointing reading the frontward/backward the back of it toward the mirror, would be the reverseof the normal reading direction. direction solid mentioned above, one would For instance,for the four-row row. So, and then the top front top back row first, have to read the matter in our world were of this sort and if it were if printed natural to look at such a printed object in the mirrorin the way described,then we would have apparent (thoughnot actual) frontOf ward/backwardreversalr. course, the actual world is not such frontward/backward. do a world. So mirrors not appear to reverser worried about why mirrors... I conclude that those who have have probablynot often(in so doing) worriedabout interpretation R or G.
III
... of Anotherinterpretation "Why do mirrors ?" is: Why is it that, when I am facing my image in a mirror,up for my image is the same directionas up for me while rightfor my image is the same D" (for direction as left for me? I shall call this "interpretation "direction").What is the puzzle that motivatesthis question? Suppose that I am standing facing a mirror.Now imagine a throughmy body ear to ear parallel to the plane passed vertically plane of the mirror,i.e., a plane that separates me into a front and a back half. Call the horizontal axis in this plane the x axis and the vertical axis the y axis. Call the perpendicular to the in the which intersects x and y axis (say somewhere the cenmirror an ter of my totso) the z axis. I shall say that rightspecifies end of
266
MIRROPs
tt1w~~~~~~~~~~~~_
x
FIGURE 3.
the x axis, up specifies end of the y axis, and frontward an specifies an end of the z axis. Now the puzzle is that, although the mirrorshould treat the x axis and the y axis in the same way, it apparentlytreatsthem differently; reversesright/left, way of specifying it a ends of the x axis, but it does not reverseup/down, a way of specifying ends of the y axis. Of course the mirroralso reversesfrontward/backward (since frontward me is the direction that is backward for my for image). But thisis not puzzling,since the z axis is perpendicularto the mirror, while the x and y axes are parallel to it. The puzzling thing is that the mirrorapparentlytreatsthe x and y axes differently,even though they are both parallel to the surface of the mirror.Notice that thereis no rotationhere-as therewas in the two puzzles discussedabove. So we seem to have a different sort of puzzle. What could account forthisdifference betweenthe x and y axes? Perhaps an anisotropy due to the earth's magneticor gravitational field,or some anisotropyin the mirror?It is easy to satisfyoneself that no such account will do, by noting that one could rotate the mirror or rotate the mirror-person systemin any way one
BUT NOT UP/DOWN?
267
chooses without changing the fact that the mirrorreversesright/ leftbut not up/down-so long as the person continues to face the mirrorand so long as nothingcomes between the mirrorand the person. The principles of simple geometricaloptics which determine that the mirrorproduces the image it produces depend only on the geometricalrelation between the surfaceof the mirrorand the reflected object. (I assume sufficient illumination.)But the laws of geometricaloptics do not treat the x axis any differently from the y axis-so we cannot appeal to them to solve the puzzle. In looking for an explanation of the phenomenon,we are in a way lookingforthewrongthing.Geometricaloptics tellsus whythe image is what it is; but geometricaloptics will not answer our question. Given a geometrical-optics account of why the image is what it is, one can reasonablyreply: "Now I understandhow this image is produced, but tell me: why is this image reversedright/ left but not up/down?" What we need is an analysis of the key termsin the question: 'up' and 'right'.The reason the mirrorappears to treatthe x axis differently fromthe y axis is that we have takenright/left a specification the ends of the x axis while takas of ing up/down as a specification the ends of the y axis. The exof planation for why the mirrortreatsup and right differently to is be found by examiningup and right,not the mirror. Up and right are quite different sortsof direction.To see this,note that 'up', on at least one standardusage, is relative7 to the earth.For example, a definitionof 'up' useful on our planet might be: away from the centerof theearth.(I shall use 'up' in thissense.) Right, on the other hand, is not relative to the earth nor to the directionof gravity, but it is relativeto the object whose rightis in question. Thus, up is the same for my mirrorimage and for me, since we are in roughlythe same spot on the surfaceof the earth, but rightis different my image than for me because right defor pends on the orientationof the object whose rightis in question, and myimage and I have different orientations: face each other. we One way of seeing thispoint: Suppose I drill a hole to the center of the earth and place a mirrortherefacingthe mouth of the hole. I then lie down over the hole facingthe mirror. The image I see is
7 In saying that a direction,x, is relative to y, I mean that y (or the direction specified by y and some other contextually determined object) is one of the factorsthat determine what direction x is. Up is also relative to a contextually indicated location, usually the location of the speaker. So a more complete definitionof 'up' in the sense indicated might be: the direction of the arrow whose head is placed at the location of the speaker and whose tail points toward the center of the earth. Of course this definitionwould not be useful in space.
268
I JROR
UPFOAGE
0
HOLE
UPFRERSON
IMAGE
PERSON
EARTH
FIGURE4.
images located at a point on the otherside of the earth,since mirror are located a distancebehind the mirrorequal to the distancefrom object. So the directionthat is up formy the mirrorto the reflected image is the directionthat is down forme.8 So with the mirrorin this rather unusual position, it reversesup/down. And, since it both up/downand right/left. it right/left, now reverses stillreverses Consider the direction clockward,the direction of the nearest (and if it is sufficiently clock.If thereis onlyone clock in thevicinity be about the same direction for both far away), clockward will won't reverseclockward.But myimage and me. So then the mirror behind my image (behind the if there is one clock immediately behind me, the mirrorwill mirror)and anotherclock immediately directionfromhim to the clock nearest reverseclockward.For the him will be the opposite of the direction from me to the clock nearestme. The general point: whethera directionis reversedby a mirror dependson what sortof directionit is, and (usually) on the as physical circumstances well. Anotherapproach to the point: imagine definingan end of the x axis (call it "right*") which is analogous to up as a specification of an end of the y axis. Define "right*" as North. Assumingmy and that we are not verynear either image and I are close together of the poles, right* for me is (approximately)the same direction as right* for my image, just as up for me is the same direction nor neitherright*/left* as up formyimage. So the mirrorreverses up/down.The point is that,if the ends of the x axis and the y axis analogous ways, there is no asymmetry are specifiedin sufficiently treatsthe twoaxes. in theway themirror to Perhaps theprecedingis not quite sufficient dissolvethe puzzle. Perhaps what is needed is: (1) a discussion of just what kind of
8 Note thatI am using'up' in the sensein whichup is thesamedirection for of at everyone a locationregardless his orientation.
BUT NOT UP/DOWN?
269
of directionrightis; (2) based on thisdiscussion,a specification the ends of the y axis, up* and down*, where up* and down* of are directions the same sortas rightand left; (3) a proofthat the both up*/down* and reverses up*/down*,and thusreverses mirror of This would provideanotherillustration myclaim that, right/left. analwhen the ends of the x and y axes are specifiedin sufficiently in ogous ways, there will be no asymmetry the way the mirror treats the two axes. This discussion will be postponed, however, form of our original question can be because a slightlydifferent introducedwhich avoids the dissolutionsketchedin the preceding two paragraphs. Call 'headward' the directionspecifiedby the vector that coincides with the vertical axis of a man's body, aimed fromhis feet and not reverseright/left toward his head. Why, then,do mirrors (I am counting this as a variant of interprehleadward/footward? tationD.)9 Headward, unlike up, is relative to the orientation of iny body. Indeed, both my rightand my headward are relative to the orientationof mybody. Yet if I am facingmy image,whereasright forme is leftformyimage,headward forhim is the same as headward for me. That is, in the sense of 'reversal'discussed here (rebut right/left not headward/footward. versald),the mirrorreverses presentedearlierwon't workhere,because both Yet the dissolution headward and rightare relativeto the orientationof mybody (and neitheris relative to the earth or to gravity).Nonetheless,we can easily dissolve this puzzle as beforeby noting that rightand headeven thoughboth are relativeto sortsof direction, ward are different of Unlike a definition 'right',a definithe orientationof mybody.10 of tion of 'headward' would referto anatomical features bodies. An of analogous specification an end of the x axis would be: left**=
'headward' by different 9 An inessentially puzzlecan be formulated replacing to the by 'topward'-where top of my body is understood be that part of my head and top up. bodywhichis usuallyor normally For peoplewho walkerect, the whereas referdetermined, of The reference 'head' is anatomically coincide. of The discussion the two in ence of 'top' dependson a regularity behavior. on with a man standing the puzzleswould be analogous.If we had started the it of mirror instead facing we wouldbe discussing pair-chest,front-instead of thepair-head, top. to applying a In ambiguous. one senseit is an adjective ?10'Right'is actually side of mybodyand someof the organson thatside. In the othersenseit is a axis to my right(in the first my from vertical to noun referring the direction is Rightwardsense) side. One mightexpressthe lattersense as "rightward." here. thesenseintended
270
the direction in the xy plane fromthe y axis toward the heart.,' Now right"* is the same directionforme as formy image; likewise for headward. So the mirrorreversesneither right**/left** nor headward/footward. Again we see that,when we have x-axis directions and y-axis directionswhich are sufficiently similar, there is no x-yasymmetry. Right*" was the same sort of directionas headward; now let us tryto framea y-axisdirectionthat is similar to right,call it "headward*." If we succeed, the mirrorwill reverseboth right/left and headward*/footward*, therebyagain illustratingmy claim that, when x- and y-axisdirections specified similarways,thereis no are in x-y asymmetry. First,we must tryto framea definition 'right'.Let us first of investigatewhetherwe can base a definitionon our way of telling which side is the rightside of an object. Such a tack oftenproduces at least prima facie plausible definitions "observational"notions. of How does one tell the rightside of an object? Here is how it seems to me thatI do it. I assign to the object in question a bottomand a front.Then I orientthe thingor myself my imaginationso that in its frontpoints in the directionmy frontpoints in, and so that its bottom points in the directionof my feet. Then the side of the object on my rightis the rightside. But how do I know which of my sides is my rightside? In my case, I just know. Many people are like me in feeling that they tell their right from left automatically. For many others, this process does not seem at all automatic. Many people utilize some anatomical or behavioral asymmetry betweentheirrightand leftsides. For example,a person may pick out his righthand as the hand that naturallygrasps a pen, or the hand with the ring or birthmark. However theydo it, most people seem to tell theirown rightsides in a way quite differentfromthe way theytell the rightsides of objects other than themselves. Thus thereseem to be two ordinaryclasses of ways of telling which side is the right side of an object. One probably learnsone's own rightside by an ostension, e.g.,by being told which side is one's right,and then one tells one's rightfromone's left by some method that depends on such a past ostension.On the other
11If you do not approve of talk of my image's heart, imagine this discussion taking place while I am undergoingopen-heart surgery. Actually,nothing in this paper depends on assuming that images have characteristics usually ascribed only to physical objects. All the puzzles can be described in termsof objects reflected in the mathematical sense rather than mirror images. (x is a reflectionof y with respect to a plane P ifffor everypoint of x there is a point of y equidistant fromP but on the other side, and vice versa.)
BUT NOT UP/DOWN?
27I
hand, one tells the right side of objects other than oneself by a process that depends on already knowingwhich side is one's own rightside. If thesejudgmentsabout ordinaryways of tellingrightfromleft definithen no satisfactory (based on informalsurveys)are correct, tion can be based on these ordinaryways of telling. For there are two classes of ways. Ways of telling right from left for objects other than myself won't do, since theypresupposea way of telling my right.And ways of telling my rightwon't do, since these ways depend on my special features. But my way of telling the right from the left of objects other based on a random choice of an than myself suggestsa definition e.g., asymmetrical object (one withno plane of symmetry), the island of Manhattan. The followingdefinitionwill serve to specifythe right side of an arbitraryobject, 0, so long as 0 is sufficiently small 12 and has a bottom(or a top) and a front(or a back). Here is the definition: The rightside of 0 = the side facingthe East River 0 is on 42nd Streetwith its bottomfacingManhattan's Drive when bedrockand its frontfacingHarlem.18. My purpose in defining 'right'was to allow me to definea y-axis similar to rightto allow direction,headward*,which is sufficiently will reverseboth right/left headand a demonstration thata mirror Since we have already seen that, given the suffiward*/footward*. * ciently similar notions right* and headward, mirrors reverse neither,we will have a convincing argument that the fact that is mirrorsreverseright/left but not headward/footward due to a difference between the two typesof direction. I shall define'headward*' as follows: my headward* is the side facing the sky when I am on 42nd Street with my front facing Harlem and myheartfacingthe West Side Highway.Now it is easy
to 12 It is a simplematter to generalize definition apply to an objectof the Let a size. arbitrary Simply superimpose set of axes on Manhattan. 42nd Street at to the be thex axis; FifthAvenue, y axis; and the perpendicular the surface a 42nd Street and FifthAvenuethe z axis. Also superimpose set of axes on 0, and one end of the y axis at the withone end of the z axis at the bottom, the The definition: rightside of 0 = thatside of 0 whichcontainsthe front. as in end of thex axis thatpoints thesamedirection theEast RiverDriveend of x Manhattan's axis when the z and y axes of the two objectsare aligned as in bottom end of O's z axis pointing the same direction the bedrock follows: as z end in end of Manhattan's axis; front of O's y axis pointing thesamedirection as the Harlemend of Manhatttan's axis. y cannotpretendto capturethe "meaning"of 130bviously such a definition thatconnects in withwhatpeoplehave in 'right' anysenseof 'meaning' meaning is mind when theyuse the word.This definition betterviewedas a "rational of reconstruction" the notion.
272
will reverse For, since to see that the mirror headward*/footward*. in mymirror image has his hearton his right, orderforhim to have his fronttoward Harlem while his heart is toward the West Side Highway, he would have to be upside down; thus headward and headward* coincide forme but not forhim; so headward* forhim is the direction that is footward*for me. Thus we have framed for similar to right/left the mirrorto a y-axisdirectionsufficiently reverse both. of A quite different interpretation "Why do mirrors... ?" is: Why is it that when I wigglemy rightarm my mirrorimage wiggleshis leftarm even thoughwhen I wigglemy head my image wiggleshis W" ('W' for head and not his feet?Let us call this "interpretation W "wiggles").Interpretation mightseem to be simplyanothervariant of the sortof interpretation (interpretation discussedin the D) previous section. But this is not so. Notice that, when I face the That is, the direcmirror, mirror the reversesd frontward/backward. backwardformyimage. But mirrors tion thatis frontward me is for do not reversew For frontward/backward. when I wiggle my front, my image also wiggleshis front.Since mirrors reversed frontward/ it backward but don't reversew frontward/backward, follows that -c reversald reversal,. Anotherproofof this point: If I turn 900 so thatmyleftshoulderfacesthe mirror, directionthat is myright the is the same directionthat is my image's right; so, in this case, the mirror does not reversed right/left. if in thiscircumstance wigBut I gle my left arm, the mirrorimage wiggles his right arm; so the mirror does reversew right/left. Once again, it followsthat reversald 7 reversal,. Nonethelessit mayseem thatthe resolutionof the puzzle given in the last section also applies here. Afterall, my right arm is just the arm in the directionthat is rightward me. And my image's for left arm is just the arm in the direction that is leftwardfor my image. Since (assumingI am facing the mirror)the direction that is rightforme is the directionthat is leftformy image (reversald), my image's left arm is opposite my rightarm; so of course his left arm wiggleswhen my right arm wiggles.Thus, whateverexplains reversald explains reversal,as well. To see that this reasoningis oversimple, note the followingfact. A suitablycurvedmirror(see Figure 5) has the property that,when I move myrightarm,myimage in thatmirror moves his rightarm. Yet the directionthat is my rightis nonethelesshis left.Thus we have another proof that reversald reversal,: right/left reversald 7/
IV
BUT NOT UP/DOWN?
273
FIGURE
5. After the figure on page 14 of Martin Gardner's The Ambidextrous Universe (New York: Mentor, 1969).
occurs with both flat and curved mirrors, but right/left reversal, requires flatmirrors. a mirrorneed do to reversed right/left All me is produce a right-side-up image facingme. The curved mirror(in the proper orientation)does this as well as the flatmirror.But, in orderto reversew right/left, mirror me the mustdo something more: it must produce an image that moves its leftarm when I move my rightarm. But the suggestionof the paragraph beforelast is roughlyright. What it leaves out is: optics. Geometricaloptics dictatesthat,when I face a (flat)mirror and move part of mybody,the part opposite it in the mirrormoves; but, since 'right'and 'head' are different sorts of notions,the part opposite my rightarm is my image's left arm, even though the part opposite my head is my image's head. So, to answerquestion W, one needs the same considerations for quesas tion D + optics. Perhaps anotherexample will clarify role of optics in answerthe ing questions like W. Consider the followingquestion: why is it that,when I point up, my image points up too, but, when I point to myright, image points to his left?The answer:optics dictates my that, when I point in a plane parallel to that of the mirror,my
274
image points in exactly the same direction-as specified by the angle with respect to a fixed set of axes. But because up and right are the (different) sorts of directions they are, up is the same direction for both me and my image, whereas right for me is left for my image (reversald).
v
All the published discussions I have seen of our question (Why do mirrors ... ?) either are limited to version G or else confusedly discuss one or another version simultaneously without distinguishing among them. For example, in his discussion of the problem in The Ambidextrous Universe Martin Gardner says: Curiously,the answer depends on the fact that our bodies, like the bodies of most animals, have only one plane of symmetry...We deone scribethe reversalas a left-right because it is the most convenient figurefrom a for distinguishing bilaterally symmetrical terminology its enantiomorph.In a strictmathematicalsense, the mirrorhas not reversedleftand rightat all, it has reversedfrontand back [emphasis added] .... We can summarizeit this way. A mirror,as you face it, shows absolutelyno preferencefor left and rightas against up and of point down [emphasisadded]. It does reversethe structure a figure for point along the axis perpendicularto the mirror.Such a reversal figureto its enantiomorph.Beautomaticallychanges an asymmetric we cause we ourselvesare bilaterallysymmetrical, findit convenientto reversal.It is just a manner of speaking, a concall this a left-right ventionin theuse ofwords.(29-31). Jonathan Bennett in his "The Left" 14 says Difference between Right and
Failure to grasp the conventionsunderlyingour use of "left" and "right"has generatedthe mildlyfamous "mirrorproblem": why does a mirror reverse left/rightbut not up/down? Martin Gardner (pp. 29-32) presentsthe only clear account I know of the solution to ... this: the answer to "Why does a mirror etc.?" is It doesn't! Your of is image in a normal mirror a visual representation an incongruous counterpartof your body, and we conventionallydescribe this sort reversal."But this conventiondoes not of relationshipas a "left/right pick out one dimension as privileged over the other two: it is merely a natural and convenient way of expressing the fact of in enantiomorphism a case where each memberof the enantiomorphic pair has-like a normal human body-a superficialover-all bilateral (Of course an object which was preciselyand totally bisymmetry. could not have an enantiomorph.)If we are to laterallysymmetrical
14
American Philosophical Quarterly,vii, 3 (July 1970): 175-191.
BUT NOT UP/DOWN?
275
describe mirror what an ordinary does,in a way whichreallydoes selectone axis of the bodyin preference the othertwo,thenwe to if must this:ifyoufacethemirror, reverses back/front;you say it you standside-on it,it reverses left/right;youstandon it,it reto if you verses you up/down. These facts, once theyare properly described, do not offer problem. a They are explainedby routineoptics.For somedeeperaspects thismatter, thepaperby Pears(181). of see What versionof "Why do mirrors . ?" could Gardner and Ben.. netthave in mind?They insistthatmirrors reversefrontward/backward (when you face the mirror).But as I pointed out (end of sec. ii, and beginningof sec. Iv) it is not the case that mirrors either reverser reverseg reversew or or frontward/backward. do mirrors Nor even appear to reversefrontward/backward sensesR, G, or W. in So it would seem that Gardner and Bennett cannot be using 'reversal' in sense R, G, or W. That leaves D. But Gardner and Bennett also insist that mirrorsreverse right/left when you stand shoulderto the mirror and up/downwhen you stand on the mirror. See the fourth sentencefromthe end in the quotation (above) from Bennett.Gardner is adamant about this point. He says: ... themirror has reversed ... and back.., execute right front a face and standfacing east,yourleftside touching mirror. before, the As the mirror reverses onlyalong the axis perpendicular it. Because to of thewayyouare standing, is now in truth this yourleft-right axis. Now you can say,in a strict geometrical sense,thatthe mirror has reversed left your and right sides, leaving unaltered your up-down and front-back axes. Imaginea mirror the ceilingor on the floor. on Again,as always, mirror the reverses onlythe axis at right anglesto itssurface (30/31). However, mirrorsdo not reversed you right/left when you stand shoulder to the mirror. Right foryou is the same directionin this case as rightfor your image. In sum, Gardner and Bennett insist: 1. Mirrorsreversefrontward/backward when you face them. reverse whenyou stand shoulderto them. 2. Mirrors right/left reverse 3. Mirrors up/down (theymean something like headward/ when you standon them. footward) Sense D makes 1 and 3 true; sense W makes 2 true,but I cannot think of any reasonable sense of 'reversal'which makes all three claims come out true and which yields a reasonable interpretation of "Why do mirrors.. . ?" I do not claim to have thoughtof every possible sense of reversal(nor have I mentionedeverysense I have thoughtof), but if Gardner and Bennett had some odd sense of
276
'reversal'in mind theysurelywould have told us about it. Moreover,I suspectthat what Gardner and Bennett do have in mind is the following: Suppose arrow A is parallel to the surface of the mirror, and arrow B is perpendicularto the surfaceof the mirror. ArrowB will be reversedin the sense that its image will point in the opposite directionfromit (relativeto a fixedset of axes). The image of arrowA, on theotherhand, will point in thesame direction could be said to reverseneither as arrowA. In thissense,the mirror the x nor the y axis, but only the z axis. (See Figure 3.) But, if this is the solution,what is the problem?Nothing said in the last five sentencesyields a sense of 'reverse'on which our question "Why do mirrors but reverseright/left not up/down?" is of any interest at all. No one in possessionof his senseswould ask: why does the image of a right-pointing arrow (parallel to the surface of the mirror)point left, whereasthe image of an up-pointing arrowpoints up too? For it is manifestly false that the image of a right-pointing arrowpointsleft.15 Moreover, even if Bennett and Gardner had had a reasonable senseof 'reversal'in mind,their'resolution'of the originalquestion would be very odd. They say mirrorsdo not reverse right/left when one faces the mirrorbecause "right/left reversal" is a "conventional description."But then theygo on to say that,if I stand withmyshoulderto the mirror, real (presumably non-conventional) right/left reversaloccurs.But whyshould it be that the description "right/left reversal" is conventional (and false) as applied to a person facingthe mirror, but nonconventional(and true) when he turns90 degrees? Massachusetts Instituteof Technology
N. J. BLOCK
15 There are other senses of 'reverse' on which the mirror could be said to reverseonly the z axis-but these senses of 'reverse' also fail to give rise to any version of our puzzle. For example, I pointed out in discussion of the Gettingversion that, if I am facing the mirrorand I superimpose myinto-the-portrait self on my image without rotation, I fail to "fit" only frontward/backward. If my shoulder is facing the mirror,I fail to "fit" only right/left, and so on. Thus we have a sense of 'reverse' (the mirrorreversesme x/y= When I superimpose myselfon the image without rotation, the image fails to fit x/y) on which the mirror could be said to reverseonly the z axis. But notice that, on this sense of 'reverse',no one could reasonably wonder why mirrorsreverse right/left but not up/down in the first place.
SUBSTANTIVIALITY
OF SPACE
277 RIGHT/LEFT
INTERPRETATIONS OF "WHY DO MIRRORS REVERSE BUT NOT UP/DOWN?" DISCUSSED IN THIS PAPER
(R) Reading Interpretation Why is it that while both the placard and its mirrorimage read up to down, the placard reads leftto rightwhile the mirror image of the placard reads rightto left? (G) Getting-into-the-portrait Interpretation Why is it that when I get into my portraitit fitsvis-a-vis but top/bottom failsto fit vis-'a-vis right/left? (D) DirectionInterpretation Why is it that when I face my image in a mirrorthe directionthat is up (headward) for my image is the same directionas up (headward) for me, while the directionthat is rightfor my image is the directionthat is left forme? (W) Wiggling Interpretation Why is it that when I wigglemy rightarm my mirrorwiggleshis left arm even thoughwhen I wigglemyhead myimage wiggleshis head too? The mirrorreversesr object x/y= The object reads y to x while the the mirror image of the object reads x to y. The mirrorreversesg x/y= When I get into my portraitit fails to fit me vis-a?-vis x/y. The mirrorreversesd = The direction that is x for my image is the x/y directionthat is y forme and vice versa. The mirror reverses,me x/y= When I wigglemyx part my image wiggles his y part and vice versa. For purposes of intelligibility, these interpretations and corresponding senses of 'reverse' have been simplified;in no case is question X really equivalent to "Why do mirrors but reverse.right/left not up/down?"
CORRESPONDING SENSES OF 'REVERSE
INCONGRUOUS COUNTERPARTS, INTRINSIC FEATURES AND THE SUBSTANTIVIALITY OF SPACE intuition,fromthe existenceof incongruouscounterparts (such as rightand left-handed glovesotherwise alike) to the existenceof space as an entityover and above the material objects in it and their spatial relations to one another. Peter Remnant and JohnEarman have argued that Kant's argument incoherent., is
1 Remnant, "Incongruent Counterpartsand Absolute Space," Mind, n.s., LXXII, 287 (July 1973): 393-399. Earman, "Kant, Incongruous Counterparts, and the Nature of Space and Space-Time," Ratio, xiii, 1 (June 1971): 1-18; parenthetical page references Earman are to this paper. to
ANT argued, partofhisargument spaceis an a priori as that

Ned Block - Why Do Mirrors Reverse Right/Left But Not Up/Down

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ned Block - Why Do Mirrors Reverse Right/Left But Not Up/Down

Uploaded by

Copyright:

Available Formats

Journal of Philosophy, Inc.

THE JOURNALOF PHILOSOPHY

Tq33J\TH3Ih5 aHMWIAV3 Oa 25OA5105IIM YHW

WWOU\qU TOVI TUff

knowjust whatquestion EOPLE who ask thisquestion rarely

WHY MIRRORS REVERSE RIGHT/LEFT DOBUTNOT UP/DOWN?

Why is it that the mirrorimage of this placard, viz., B, reads from

oa 2lOHflIM YHW 3fl3V3R TI31\THDIfl NWO\q9U TUO TON

THE JOURNAL OF PHILOSOPHY

MHA WIHBO2 2E ICHH\rEL DO IHEAEBE

Doli nb\DOMNS mO1

1. Viewof theroomfrom side. thie

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

/\6AtAFTp 8UT/OT MR~\/DOW?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

American Philosophical Quarterly,vii, 3 (July 1970): 175-191.

WHY DO MIRRORS REVERSE RIGHT/LEFT

BUT NOT UP/DOWN?

THE JOURNAL OF PHILOSOPHY

ANT argued, partofhisargument spaceis an a priori as that

You might also like