Lay people like to think that contracts have to be in writing. This is only true in specific situations and even then it’s not entirely correct. A set of rules known collectively as the Statute of Frauds makes it necessary to have some things in writing if you want to enforce the contract. This requirement only applies in certain situations, like the sale of land or goods that are over $500 in value. In those cases, you don’t really need “the” contract in writing, but just to have some documentation with enough details that links to the transaction.
The idea is to make fraud marginally harder. If someone claims you promised to sell them your house, and you say that they are lying, we don’t want to just make the case depend on who the jury find more believable. We can save a lot of trouble by asking that the alleged buyer, as a threshold, show evidence *in writing* that he bought the house.
Critically, the writing has to involve the signature of the other party. The alleged buyer should be able to show that you actually signed whatever document they managed to produce. If you think that people lying about non-existent transactions is a larger issue than people getting out of valid transactions on a technicality like a signature, then this rule makes sense.
Recently, Professor Dave Hoffman noted (not linking because he auto-deletes his tweets, which I take as a request not to post screenshots) that according to some court decisions, signatures are also needed in *text messages*. I think this requirement is a mistake. This is an important topic, so I’m going to give a review of the doctrinal landscape as I see it and then the policy arguments.
I. The Common Law: From telegrams to text messages
Let’s look back at how the common law dealt with two other disruptive technologies — the telegram and the fax. Do names types at the bottom of telegrams satisfy the SoF? Yes (Lamle v. Mattel, Inc., 394 F.3d 1355, 1362 (Fed. Cir. 2005)). But it was necessary to have this signature someplace (McNear v. Petroleum Exp. Corp., 208 Cal. 162, 166, 280 P. 684, 686 (1929)). On the other hand, Fax headers were not considered enough. (Parma Tile Mosaic & Marble Co. v. Est. of Short, 87 N.Y.2d 524, 526, 663 N.E.2d 633, 634 (1996)). The problem with this case is that it’s from the pre-digital era. Modern courts seem to disavow it (“we find that Parma is not controlling.” Philadelphia Ins. Indem. Co. v. Kendall, 197 A.D.3d 75, 78, 151 N.Y.S.3d 392, 395 (2021)).
Transitioning to the digital age, the states have adopted the Uniform Electronic Transactions Act (UETA). It recognizes the validity of electronic signatures. (UETA doesn’t apply to every transaction, but I’ll abstract from this consideration). Thus, having an electronic signature — including in a text message–should be enough.
The problem is that UETA does not define the term “sign.” It does, however, provide some advisory guidance in the comments: “One may use a digital signature with the requisite intention, or one may use the private key solely as an access device with no intention to sign, or otherwise accomplish a legally binding act. In any case the critical element is the intention to execute or adopt the sound or symbol or process for the purpose of signing the related record.” In other words, what matters is whether we can infer from the communication an intention to “execute or adopt” the record. [Unif. Elecs. Transactions Act § 2 ]
The comments further clarify that electronic signature is attributable to a person even when generated by a machine. “so long as the electronic record or electronic signature resulted from a person’s action it will be attributed to that person …. [a] person’s actions include actions taken by . . . an electronic agent, i.e., the tool, of the person.” [Unif. Elecs. Transactions Act § 9 ].
UETA opens the possibility that text messages will be self-signing. The clearest example will be an automated tool that adds “Signed by Yonathan Arbel” to each one of my texts. Aside from alienating anyone I text with, it would have the effect of actions taken by “an electronic agent.” Indeed, one Texas court decided that automatically attached signature blocks in emails are the same as manual digital signatures. (“Considering a signature block an electronic signature is thus in line with . . . UETA’s expansive purposes and rationales.” Williamson v. Bank of New York Mellon, 947 F. Supp. 2d 704, 711 (N.D. Tex. 2013)
How far can we push it? Another Texas court thought the answer is very far. Take an email that has no signature, automatic or otherwise, but has the sender’s name in the FROM header? The court said that’s kosher. (“The name or email address in a “from” field functions as a signature in an email. Khoury v. Tomlinson, 518 S.W.3d 568, 577 (Tex. App. 2017). Other jurisdictions adopted a similar approach (Int’l Casings Grp., Inc. v. Premium Standard Farms, Inc., 358 F.Supp.2d 863, 873 (W.D. Mo. 2005); Kluver v. PPL Mont., LLC, 368 Mont. 101, 293 P.3d 817, 822–23 (2012); Dalos v. Novaheadinc, No. 1 CA-CV 07-0459, 2008 WL 4182996. Nesbitt v. City of Bullhead City, No. CV-18-08354-PCT-DJH, 2020 WL 6262396, slip op. at *4 (D. Ariz. Oct. 23, 2020).
Not all courts are equally lenient. A recent case from the Supreme Court of Mississippi adds a bit of interesting nuance. In Parish Transport, an email regarding a $1.25 million equipment purchase was sent by one of the parties. The seller responded with the colorful “Got to talk to my partner before we take a mule kick in the nuts on this deal. ” This email had what appears to be an auto-signature, with the seller’s name and contact info. After talking to his partner, the seller sent an email saying “Ok. Let’s do it.” alongside the familiar “Sent from my iPhone.” The parties continued correspondence, but no more signatures were found. Eventually, the seller got a better offer and wanted to back out. Ironically, his email retracting his agreement DID include his signature.
The court decided that, under UETA, there is no issue with emails satisfying the Statute of Frauds, citing Preston Law Firm, L.L.C. v. Mariner Health Care Mgmt. Co., 622 F.3d 384, 391 (5th Cir. 2010) (“Emails can qualify as the signed writings needed to form contracts.”)
The court refused to follow, however, the Khoury case from Texas, saying that Mississippi requires a “different level of intent,” not just an intent to sign the document but an intent to “accept or adopt the writing” itself.
In this case, there were (possibly automatic) signatures in some of the emails, but not in the relevant email. The court invokes the composite document rule and says that it is possible to treat the entire email chain as one big document. But whether or not the signature was “intended to adopt the entire email chain as a whole,” is a factual question, best left to the fact finder. Other courts also subscribe to this stronger requirement that the intent will be to adopt the document rather than to sign the document. So we have a split.
OK, enough emails. What about text messages? It’s not a big surprise that there isn’t a whole lot of caselaw on the issue (yet), but there’s some. The emerging rule seems to be, as Prof. Hoffman noted, that signatures are required. In addition to his examples, others include BrewFab, LLC v. 3 Delta, Inc., No. 8:20-CV-2031-VMC-SPF, 2022 WL 138095, at 7 (M.D. Fla. Jan. 14, 2022); Fiore v. Lindsey, No. 17 MISC 000533 (RBF), 2017 WL 5969332, at *4 (Mass. Land Ct. Nov. 29, 2017), judgment entered, No. 17 MISC 000533 (RBF), 2017 WL 5907443 (Mass. Land Ct. Nov. 29, 2017); Walsh v. Abate, 336 So. 3d 50, 53 (Fla. Dist. Ct. App. 2022).
II. What Should Courts Do?
How do you reconcile Kluver (“from” is enough”) with St. John’s (unsigned text message isn’t enough)? I don’t think you can or should.
As a preliminary manner, this rule is not very effective from a pragmatic perspective. I don’t like arguing from expectations, but the signing-texts-rule is extremely unnatural. I think Hoffman was alluding to the strangeness of people signing their own text messages. It’s also not great to have medium-specific rules, with one esoteric rule for emails and another to text-message. Of course, we could just require signatures in emails as well, but that only leads to the core issue: we don’t need this rule.
Here’s the key point: ALL text messages already have a signature. So do emails. It’s just the the act of signing is not what we expect from paper-era documents. Rather than an illegible scribble in faux cursive, text messages use a different mode of signing.
What is a signature? It is a method of (1) authenticating one’s identity and (2) a symbol that one endorses the appended text. If one adopts a Fullerian perspective, signatures can be parsed as producing evidence, channeling behavior, and cautioning parties about the significance of their actions (for a development of these ideas, see Michael J. Hays, The E-Sign Act of 2000: The Triumph of Function over Form in American Contract Law, 76 Notre Dame L. Rev. 1183, 1206 (2001))
On the authentication front, typed signatures are unnecessary in the digital age. Both emails and text messages contain more information than is visible on your device. Both contain “meta data.” The data is technical, but it includes information about the sender (see some background here). The existence of the data makes it difficult to draw a meaningful distinction between the from header in an email and the phone number of the sender of a text message. More substantively, both indicators offer a verification of sender identity. They are not perfect, but neither are wet ink signatures. In fact, the moment we accepted typed signatures, we have conceded this ground.
What about endorsement? Here, we do want to make sure that the communicator indeed accepts the content of the message. In the paperverse, it’s hard, because documents can be produced unilaterally. It is hard to know whether a document–much less a printed one–is the product of collaboration or not. So we need evidence of collaboration–we need an endorsement by both parties. We don’t require much; initials are fine (Sanborn v. Flagler, 91 Mass. 474, 478 (1864)). The point is simply to
With text messages, however, the problem is almost trivial. The moment a person hits ‘send’ they produce the digital paper-trail that evidences that the written document is the product of collaboration.
If one is willing to accept automatic signatures as valid signatures, it’s hard to see how this is any different. In both cases, the act of hitting send accomplishes the same end. There is the same degree of deliberation. When an attorney hit send on a settlement offer he regretted, the court had this comforting words to say ” “if an attorney hits ‘send’ with the intent of relaying a settlement offer or acceptance, and their email account is identified in some way as their own, then it is unnecessary for them to type their own signature.” Philadelphia Ins. Indem. Co. v. Kendall, 197 A.D.3d 75, 78, 151 N.Y.S.3d 392 (2021))
A final note on the Fullerian perspective noted about. At this point in our acculturation to the digital age, hitting send on a text message is not qualitatively different than signing a document (aside from the novelty of the latter). Whatever the cautionary elements of signatures were, they seem easily replaced by the ‘send’ button.
So, overall, requiring signatures on text messages is the wrong approach. Of course, we should be worried about the intent to create a contract that is communicated via text. But this is a separate concern from the requirements of the Statute of Frauds. The SoF is very powerful–it allows parties to escape valid Ks on a technicality. We shouldn’t require parties to sign text message lest they will find themselves without recourse.