I am looking for a correct and common definition of the glass that you can see through from only one side, like those you see in detective TV series.

enter image description here

I could find the definition:

one-way mirror from the AHD :

A mirror that is reflective on one side and transparent on the other, often used in surveillance. Also called two-way mirror.

The definition corresponds to what I am looking for but I don't understand the part I put in bold.

It looks like it means that it is called both one way and two-way mirror.

Does it make sense? and could it actually create a misunderstanding? What is the appropriate, unambiguous way to call it?

  • 1
    Both terms are unambiguous. Though admittedly, having two apparently contradictory terms for the same thing is unusual. Google Ngrams indicate that 'one-way mirror' is the more common term. Aug 1, 2016 at 16:44
  • 2
    They both usually mean the same thing, but one-way mirror is much more common. I suspect some people sometimes use two-way mirror to mean a half-silvered mirror (where if light comes in at 45°, half is reflected, and half absorbed). Aug 1, 2016 at 16:49
  • 1
  • 1
    The term "one-way window" would be more correct from a technical point as you can see through the object. The term mirror suggests an ability to reflect at the expense of transmitting.
    – Stan
    Aug 1, 2016 at 18:13
  • 5
    The mirrors in question actually are equally transparent and equally reflective of light from either direction: any asymmetry in observers' experience, seeing the thing as a mirror from one side and as a window from the other, is wholly dependent on different levels of illumination. Aug 1, 2016 at 18:38

3 Answers 3


Compared to the phrases one-way street and two-way street, one-way mirror and two-way mirror don't make a lot of sense. In looking at early instances of the two mirror terms, I was surprised at how a phrase (two-way mirror) that originally had a number of different meanings seemed to gravitate toward the meaning of its n – 1 doppelgänger (one-way mirror).

Early alternative senses of 'two-way mirror'

The earliest use of either term that I've been able to finds is from an advertisement for Moore's, a clothing store, in the [Perth] West Australian (April 8, 1937):

19/6 HANDBAGS. 14/6

Imported lizard, snake and croc. Handbags. Silk lined, finished with chrom. clasp and two-way mirror. Usually up to 19/6 each: Thursday, Friday and Saturday, 14/6 each.

Here the "two-way mirror" evidently consists of two mirrors set at different angles to the viewer, or something of the sort. There is certainly no value in building a see-through mirror into a handbag. Another early match is from Hardware Age, volume 146 (1940, date not confirmed) [combined snippets]:

Attractive display material, national advertising, direct-mail folders and other dealer helps are all available through the Modern Kitchen Bureau, 420 Lexington Ave., New York City, N. Y. Window display material carries out the "Through the Looking Glass" theme of "Alice in Wonderland." This center display is 50 in. wide and 38 in. high. The two-way mirror changes, seemingly by magic, from a perfectly polished reflecting surface to a catchy colorful selling picture right before the prospect's eyes.Four 11 in. by 14 in. full-color display cards stressing the four major reasons why women want electric ranges are a part of the trim. Cost of window display with four counter cards and five range cards $3.45 per set.

The technology behind this type of two-way mirror is discussed in two later articles—one in Popular Science (1948) and one in Scholastic Voice (1955).

And from "Two-Way Mirror," in the [Charters Towers, Queensland] Northern Miner (September 6, 1943):

A novel two-way driving mirror for motor vehicles has been patented in Great Britain. Its object is to enable a driver to see not only overtaking traffic but oncoming traffic when he is about to pass in front of large vehicles moving in the same direction. It is claimed that the latter advantage reduces the risk of head-on collisions an lessens wear on tyres and brakes due to sudden "pull-outs" and "pull-ins." The device consists of two mirrors attached together by an arm which permits of adjustment of the second one. The first mirror, which is located in the normal position, is divided horizontally, the upper part showing overtaking traffic and the lower part oncoming traffic through the reflections of the second mirror. A number of bus undertakings in England is stated to be experimenting with the two-way mirror.

Early instances of 'two-way mirror' in the modern sense

By the late 1940s, however, two-way mirror was also being used to refer to a mirror that a person behind it could see through while the person facing it could see only a reflection. From Illinois Department of Public Welfare, Annual Report of the Department of Public Welfare for the Fiscal Year July 1, 1947–June 18, 1948 (1948):

A total of 222 patients was examined psychologically by intensive batteries of tests. Concentrated office and seminar space, and three examination rooms with two-way mirrors and inter-communication systems were set up, and considerable equipment was added to the laboratory facilities.

From "Claims Hotel Record," in the [Adelaide, South Australia] Mail (November 19, 1949):

In 1936, in Paris on holiday, Mr. Nimmo saw a novel idea which he installed on his return. He has two-way mirrors set in the walls of his office so he can keep an eye on the bar. Those in the bar cannot see through, even if they are close to the mirror there. Mr. Nimmo has his own repair shop complete with metal lathes and tools, where he does most hotel repairs himself. He said today he intended to keep on with the hotel until he died.

And from "Photos Taken in Nude," in the [Broken Hill, New South Wales] Barrier Miner (July 19, 1952):

Daytona Beach (Florida), Friday.—Olympic and national swimming champions were among women secretly photographed by means of a two-way mirror at the municipal swimming pool here, investigators were told today.

The assistant city manager (Mr. Edmondson) testified that about 130 photographs, showing women in various stages of undress, were found in a locked cupboard.

A two-way mirror in the women's dressing room backed up to the cupboard, so that anyone inside could look through into the dressing room, he said.

Early instances of 'one-way mirror'

Early instances of one-way mirror[s] adopt the see-through trick mirror meaning right away. From Alexandra Schachter, "The Rate of Blinking in Tension Situations" (1944):

The laboratory in which the experiment was conducted consisted of two rooms. A one-way mirror built into the wall between the rooms enabled the subject to be observed without her knowledge. During all but one of the experimental situations (including the controls) the subject was seated about 6 inches from the mirror at right angles to it. The experimenter observed the subjects from the adjoining room, by way of the one-way mirror.

And from "Through the Looking Glass," in the Goulborn [New South Wales] Evening Post (August 29, 1947):

With one of those "I-see-you, you-see-yourself" sort of magic mirrors installed in your front door; an approaching caller sees only his own reflection in the glass, but you, on the other side of the door, see through it as though it were not a mirror at all and take a good look at him without being seen. Not only are these one-way mirrors valuable in outside doors, but especially are they useful in swinging doors such as are often found between the kitchen and the dining room. Guests in the dining room cannot see into the kitchen, but the hostess or servants in the kitchen look through the "mirror" to observe the progress of the meal.


Logically, you might expect one-way mirror to refer to a normal mirror, and two-way mirror to refer to a mirror that acts as a mirror in two directions or ways. But as the terms one-way mirror and two-way mirror are popularly understood, neither term has an especially logical meaning.

It might have been more logical to call something that acts as a mirror when viewed from one side and as a window when viewed from the other a one-way window; after all, a normal window is two-way, except when darkness on one side of the glass prevents the viewer from seeing anything on the other side. But one-way window isn't the term that English speakers settled on.

In adopting the term two-way mirror, people seem to have wanted to emphasize that whereas a normal mirror offered only one viewing position (from the front), the new kind of mirror offered two (form the front, of course, but also from the back, albeit not with anything like a mirror view). Hence, two-way mirror = views in two ways, one way as a mirror and one way as a window.

Conversely, people who adopted the term one-way mirror to describe the same two-way viewing setup seem to have wanted to emphasize the idea that although there were two viewing positions for the new kind of mirror, only one was a mirroring view. Hence, one-way mirror = views in two ways, one way as a mirror and one way as a window.

So depending on how you emphasize the definition that applies to both terms, one-way mirror makes some sense and two-way mirror makes some sense. But closer scrutiny suggests that neither flawed term is overwhelmingly more logical than the other, and that may be why English continues to make room for both, as this Ngram chart comparing one way mirror (blue line) versus two way mirror (red line) versus one way mirrors (green line) versus two way mirrors (yellow line) for the period 1930–2005 illustrates:

Update (11/24/2017): A note about Ngrams of hyphenated words or phrases

A comment below from Kevin recommends the "clearer" Ngram that results from a request for an Ngram of "one-way mirror" and "two-way mirror." Here is the chart version of that Ngram graph:

The lines are indeed clear and distinct—but I challenge anyone to persuasively explain the relevance of what they show to the request for line graphs of "one-way mirror" and "two-way mirror."

The problem is that Ngram doesn't track instances of "one-way mirror" and two-way mirror" as such. Instead—as the program points out on the Ngram graph (but not on the Ngram chart)—it replaces those character strings with "one - way mirror" and "two - way mirror," respectively. Though I don't know precisely what "one - way mirror" and "two - way mirror" mean to Ngram, my sense is that the hyphen functions as a minus sign rather than as a hyphen in the replacement form. The practical reason that Ngram avoids hyphenated word forms is that its programmers wrote it in such a way that it can't distinguish between, say, "long-standing" and "longstanding" when a line break occurs after "long-".

Let's look again at Kevin's Ngram chart, but this time let's supplement its entries for "one - way mirror" and "two - way mirror" with entries for the four terms that I tracked in my version of the chart: "one way mirror, " "oneway mirror, " "two way mirror, " and "twoway mirror." Here is a link to the resulting Ngram graph (which includes links to supporting data from Google Books, beneath the graph) and here is the chart version of that graph (which omits the supporting links):

The most striking thing about this Ngram chart is how much greater both "one - way mirror" and "two - way mirror" are than any of the other four terms. But if you look at the Ngram graph shown in the accompanying link, you'll see something even more striking: whereas the Ngram page includes links to instances of Google Books matches for the four terms from my original chart, it doesn't show any links for "one - way mirror" or "two - way mirror." And if you check the links themselves, you'll see that the "one way mirror" links actually go to instances of "one-way mirror," and that the "oneway mirror" links (reachable if you click "Search instead for 'oneway mirror'" beneath the note "Showing results for one way mirror") actually go to instances of "one-way mirror" in which a line break occurs at "one-".

Clearly, the links for the four nonhyphenated forms that I specified go to hyphenated forms; and just as clearly, the line graphs for "one - way mirror" and "two - way mirror" are much greater than the sum of the line graphs for "one way mirror" + "oneway mirror" and "two way mirror" + "twoway mirror," respectively. The Ngram graph explicitly states that Ngram has replaced the terms "one-way mirror" and "two-way mirror" with the terms "one - way mirror" and "two - way mirror" and that it has done so "to match how we processed the books."

Furthermore, although the graphed frequency of either "one - way mirror" or "two - way mirror" is much greater than the sum of the four frequencies of the terms that I added to Kevin's Ngram, the program provides no link to any occurrence of either "one - way mirror" or "two - way mirror." I don't doubt that Ngram used some sort of data matching to produce those two line graphs, but I have no idea what its matching criteria were—and I doubt that anyone outside Google knows either. For practical purposes, the two line graphs in Kevin's Ngram are undefined and unsupported by tangible data.

The "Advanced Usage" section of the "About Ngram Viewer" page has this note about what it calls "Ngram Compositions":

The Ngram Viewer provides five operators that you can use to combine ngrams: +, -, /, *, and :.

- subtracts the expression on the right from the expression on the left, giving you a way to measure one ngram relative to another. Because users often want to search for hyphenated phrases, put spaces on either side of the - sign.

So aside from being unable to distinguish between hyphens as compound-word markers and hyphens as line-break signals in normally hyphenless words, Ngram uses the hyphen character as a special operator. And unfortunately, the advice to put letter spaces on either side of the - sign in order to make it function as a special operator is quite superfluous, because Ngram does that for you, automatically, whether you want it to or not.

The only other mention of hyphens in the "About Ngram Viewer" page appears in a bullet point response to a frequently asked question:

Why am I not seeing the results I expect?


  • Your phrase has a comma, plus sign, hyphen, asterisk, colon, or forward slash in it. Those have special meanings to the Ngram Viewer; see Advanced Usage. Try enclosing the phrase in square brackets (although this won't help with commas)."

We've already seen what "Advanced Usage" had to say about using a hyphen as a special operator. But the suggestion about enclosing a search term in square brackets if it contains a character that also functions as a special operator gave me hope that I might be able to specify that I wanted to track hyphenated words—that I wanted Ngram to read the - in both "one-way mirror" and "two-way mirror" as a hyphen, not a minus sign. So I asked for an Ngram of "[one-way mirror]" and "[two-way mirror]," putting two terms in square brackets, but the result was this error message:

No valid ngrams to plot!

Ngrams not found: one-way mirror, [one-way mirror], two-way mirror, [two-way mirror]

The Ngram Viewer is case sensitive. Check your capitalization!

The characters +, -, *, / require parentheses to be interpreted as a composition.

Adding parentheses around the hyphens yielded the same "No valid ngrams to plot!" message, and replacing the square brackets with parentheses caused Ngram to generate oddly flat line graphs of "(one – way mirror)" and "(two – way mirror)", along with links to matches for the phrase "way mirror" and for the words one and two. As a final tactic, I tried enclosing just the hyphens in square brackets, but one again the Ngram Viewer responded with "No valid ngrams to plot!"

My conclusion, after multiple highly unsatisfactory experiences with Ngram's handling of words and phrases containing hyphens, is that Ngram programmers never got around to making the program's "search for hyphenated phrases" feature functional. Instead, they seem to have instructed Ngram to transform requests for plots of hyphenated words into requests for plots of advanced Ngram comparisons of the component words or phrases within the originally specified hyphenated word or phrase. The upshot of all this is that I still haven't been able to find a way to get Ngram to generate meaningful line graphs of hyphenated words or phrases of the type that Kevin wanted to create.

  • +1. An alternative phrasing follows your Goulborn quote - that is, the mirror's 'image' (disregarding lateral inversion) is sent to 2 places at once - its front and back, hence a 2-way 'mirror'.
    – Lawrence
    Aug 3, 2016 at 0:01
  • 1
    The two-way mirror mentioned in the handbag advert might have been either (a) a reversible mirror with increased magnification on one side (e.g. for applying makeup) or (b) a non-reversing mirror -- two mirrors that open at 90° to show you how others see you.
    – TripeHound
    Aug 21, 2017 at 14:05
  • Hyphenation gives a much clearer Ngrams chart.
    – Kevin
    Nov 24, 2017 at 18:07
  • @Kevin: Would that it were so. Unfortunately, Ngram does not track instances of "one-way mirror" and two-way mirror" as such. Instead it replaces those character strings with "one - way mirror" and "two - way mirror," respectively. I don't know precisely what "one - way" and "two - way" mean to Ngram, but my sense is that the hyphen is functioning as a minus sign rather than as a hyphen in the replacement form. The reason Ngram avoids hyphenated word forms, I think, is that it can't distinguish between, say, "long-standing" and "longstanding" when a line break occurs after "long-".
    – Sven Yargs
    Nov 24, 2017 at 19:14
  • 1
    I asked on Webapps.SE, hopefully they will know.
    – Kevin
    Nov 25, 2017 at 4:29

Yes it makes sense.

It is actually a one way mirror as far as the common understanding of a mirror is concerned. You can see through it on one side, and it has a mirror on 'one' side.

One way mirror

noun - see this link

  1. a sheet of glass that can be seen through from one side and is a mirror on the other, used especially for observation of criminal suspects by law-enforcement officials or witnesses.

    Also called two-way mirror.

The short answer to your question, is you can indeed use the terms synonymously. But to avoid ambiguity the best choice would be "one way mirror" as this is the main entry in the dictionary for this type of mirror, and conceptually if you think of the device you are describing we know it to only have a 'mirror on one side'.

So I would refer to it as a 'one way mirror' to avoid ambiguity.

For a lengthier more technical discussion, you might check out this reddit discussion - in which they talk about the fact that both sides of the mirror are actually reflective.

That type of mirror doesn't reflect just one way - light from both sides are transmitted and reflected. It only works when there is a huge discrepancy in lighting conditions - for example, the interrogation room is brightly lit, while the observation room is dark.

The mirror works by reflecting only a percentage of light. Since there is so much light coming from the interrogation room, whatever percentage that is reflected drowns out any light that is transmitted from the observation room, so to them it looks like a mirror. ref

But from the language point of view, both 'one way mirror' and 'two way mirror' are synonymous.


I think more commonly "two-way mirror" is used. It makes sense to think of the common mirror as a simple "one-way mirror". That makes some sense scientifically as the photons radiate off your body, then bounce off the mirror and into your eye. The photons are going one way (with a bounce).

In the Interrogation Room the photons radiate off the suspect's body, then some bounce back to the suspect, but others pass through the glass to reach the observer's eyes. The photons are going two ways.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.