Forensic Linguistics: Language Analysis in Criminal Investigations

[00:00:00]Welcome to our Hofstra University virtual zoom event. I'm Rob Leonard, and I'm going to be talking about forensic linguistics Whenever we deal with crime, unfortunately, there are a lot of distressing and violent and often gruesome details of crimes and I just wanted to warn people of that When I worked with the FBI Behavioral Analysis Unit one at Quantico To do training of their folks and other agents worldwide Jim Fitzgerald who had recruited me down there to help him teach These week-long Forensic linguistics boot camps had this as his mantra language can solve and prevent Crimes and as we went around and trained other people, we always brought this up Language evidence is something that is still being overlooked, but just about every time we do a presentation Someone would come and say, you know, it never occurred to us At the scene of the murder there was a letter that didn't seem to be in place and Then can you look at it?

[00:01:24]Very interesting thing about language is that every time any of us open our mouths or sit down at a keyboard We can be revealing a lot more than we're aware that we are revealing For example where you're from where you've lived how old you are etcetera case after case We've been able to ring remarkable amounts of intelligence from seemingly minor linguistic details and We'll see a couple of cases where this is what happens.

[00:02:10]Embedded in your language and your history there are clues as to what your nationality, your occupation, your education, where you've lived, as I say, what movies you've seen and skilled persons can sometimes find that information. Now I'll show you a case These are typical assignments for forensic linguists So linguistic demographic profiling, and we'll talk about the Unabomber in a minute.

[00:02:53]Authorship analysis and the worksheet that we sent before the the session has the Hummert case letters and then the spray paint on a wall here is from the Coleman triple homicides.

[00:03:19]and Discourse analysis and analyzing ambiguous phrases and underlining meaning and this is a case of Tonya Christiansen a University of Copenhagen professor with whom I work and who teaches with me here every Fall, we'll talk about that too, later.

[00:03:43]These are some of the criminal domestic foreign counterterrorism cases that we have dealt with just in the past couple of years.

[00:04:14]I am lucky to be involved with a number of organizations here at Hofstra.

[00:04:19]We have the graduate program in linguistics forensic linguistics We also have a joint masters and juris doctor program with the Hofstra Law School we have a five-year bachelor's and master's in linguistics forensic linguistics and We have our Institute for forensic linguistics Threat assessment and strategic analysis and our capital offense or death penalty Innocence Project with Hofstra Law School And I'll be talking about that more. I Also have a consulting firm and I and some others have just started what we're calling FLINT a high-tech artificial intelligence startup trying to code and create an engine that actually does what experienced forensic linguists do.

[00:05:14]Forensic linguistics More and more is being admitted and has been admitted in court cases of murder, death threats, forgery, fraud, corporate espionage, all of these because all of these deal with language and language evidence and if you think about it what in the in the law is not language? Even a forensic chemist is going to have to testify through English here in the United States or in another language. Oh I wanted to mention before That I'm really happy to have questions but I think we should hold them to the end but write them now and put them on in the chat and as they occur to you.

[00:06:01]So I was talking about being admitted to testify in court and in the United States There are two rules the Daubert and Frye which by which Judges are charged to be the gatekeepers of scientific evidence, of expert evidence, doesn't have to be scientific per se, and these are some of my background items that judges used to weigh whether or not to allow me to testify and I've been fortunate because we have I've been admitted in twelve different state courts several times and six federal district courts and I've also testified for international tribunals and I have consulted to criminal civil counterterrorism agencies all over United States, Canada England the United Kingdom Europe, Latin America, and Asia for clients including the FBI, the Joint Terrorism Task Force, and the Prime Minister of Canada.

[00:07:18]So, the idea of forensic linguistics as a useful tool is growing.

[00:07:25]Here is an example that is demonstrates what my mentor and the founder of forensic linguistics the United States named Roger Shuy, s-h-u-y, used to help in the kidnapping of a child, this is a famous example at least among forensic linguist and since we're not as well-known as a fingerprint people, it's probably not that widely known.

[00:07:58]So take a look at this ransom note.

[00:08:08]You see some things of course right off the bat and here are some Linguistic investigative features. This is a non exhaustive list, but just some of the things that linguists look at when they're trying to understand who might have written something.

[00:08:40]The most important thing here in a forensic context is beware of disinformation. Unless somebody is going to sign his or her name to the threatening letter or to the ransom note and say where they live, we have to immediately assume that there are going to be other disinformation items in there as well. So a common one, of course, is dumbing down.

[00:09:07]It's a lot easier to dumb down than it is to, what I jokingly call, dumbing up, and we've seen in countless cases where people try to dumb down.

[00:09:19]What happens, though, is that they don't dumb down all systems of language the same amount and when things don't match; So here take a look at this now.

[00:10:03]So what do we see here? Of course, we see all sorts of anomalous stuff.

[00:10:08]We see Trash Can misspelled, we see cops misspelled We see Devil Strip, whatever that means.

[00:10:17]So, we generate hypotheses. So, "trash can," okay. It's a K for the "kuh" sound, cops K for the "kuh" sound. So, maybe this person is not a native English speaker and whatever language the person has as their native language, you always use a K for the "kuh" sound.

[00:10:38]But then we have Carlson Okay, maybe he or she looked up or knows the street name is actually spelled with a C But then we have cash then we have Come alone. So, it's an irregularity, it's anomalous. If somebody really always spells it with a K, then they should always spell it with a K. And then there are other things wrong.

[00:10:59]So, for example, daughter is misspelled, as well as cops and can, but "precious" in the first line is spelled correctly, and, as a matter of fact this isn't how ransom notes usually are.

[00:11:14]Usually it's send the money leave the money or the person dies But here this use of precious Really bespeaks a native English speaker with a good command of the language or at least someone with a good command of the language because it's a turning of the knife into the hearts of the parents. "Your precious little girl." And indeed we have this "devil strip" thing, too. Now Roger, well, all of us, usually it takes us a while to figure these things out, but Roger almost immediately said to the authorities, "Do you have on your suspect list a well-educated man from Akron, Ohio?" And I always like to think that the police said, "Come on, Roger, stop joking. How could you know that and so quickly?"

[00:12:09]You know, did did he, does he walk with a limp you know Sherlock Holmes always says the man who wrote this walks with the limp, has a frayed left sleeve, and was in the British Army in India. And Roger said, "just look at your suspect list" and the answer was that they did have someone who was a well-educated man from Akron, Ohio and he turned out to be the person. Now, how did Roger know all this?

[00:12:35]Okay. Well, we've just seen all these poorly done dumbing down and also the word "devil strip" is used for this patch or a line of grass but it is used only in Akron, Ohio, and surroundings.

[00:12:59]How does one know this? Well If one is Roger Leonard, uh Roger Leonard, then if one is Roger Shuy people often call him Robert Shuy some reason So I guess I just did the same thing myself.

[00:13:15]Now, if you have a word for this strip of grass as I did growing up, you probably figured that that's the word that everybody in the world has because when do you ever talk about this strip of grass? The way I know my term in this area that I was brought up in is because my father was a county official and here it's called the county strip, and it was the part of the lawn that I hated to mow because it had all these trees on it and he was always saying you didn't mow the county strip. So, I grew up assuming that everywhere in the world it was called the county strip.

[00:13:48]But now I found that it's called some other things, and this was a very good bit of evidence for Roger because it was not the thing that somebody knows people have a lot of different words for and therefore might try to disinform by using the word that was not from his or her region.

[00:14:10]What other features show the mistakes to be disinformation and dumbing down?

[00:14:15]So, take a look now that we've seen all these other things, what else?

[00:14:24]I'll give you a hint: the punctuation.

[00:14:32]The punctuation is perfect and hardly in keeping with someone who doesn't know how to spell cops.

[00:14:42]So, what forensic linguists do is we maximize the intelligence yield from this linguistic evidence: texts, reports, wills. I've had a number of cases where wills were changed just before the person died and who benefited from the will changed mightily and wery often fraud examiners and others know that there's something wrong about the communications, but they can't put their finger on it because they have their own tool kit and we have our tool kit and in one I was able to show that the person who had changed this lady's will was not the lady at all, but her relatively new husband and he had redirected virtually all her considerable wealth to his children from an earlier marriage and away from her children from an earlier marriage.

[00:15:44]So, let's do some more linguistic demographic profiling.

[00:15:50]Everybody knows that in Spanish you put an inverted question mark. *speaking Spanish* This is so different that it's very unlikely a Spanish speaker writing in English would do it. Matter of fact, if you see it in English, you probably should think that is a a disinformation, somebody tried to frame a Spanish speaker.

[00:16:11]But there are other more subtle differences than this inverted question mark.

[00:16:18]Capitalization is different in Spanish and other languages than English.

[00:16:23]So, you say, you write *speaking Spanish* Espanol is not capitalized and that's an easy sort of subtle thing for somebody to carry over into writing English.

[00:16:38]So, in one case we saw "I'm always watching when she goes to French class" was this evidence of something or maybe it was just a typographical error or somebody was too lazy to hit the shift bar, but it merits further looking at Now in a real threat case, a stalking threat case we have this a "challenge that you have the right to have her to yourself. I have known her since a very long time myself, perhaps even longer than you."

[00:17:15]And it was something in here, that just wasn't Something I Must have read it 30 times. I didn't know what was wrong with it. My colleague and mentor from Colombia Dr. Benjy Wald said ah since then I realized English says things like I've known her since 2013, but we don't say since a very long time in English.

[00:17:54]But there are languages that do, for example French. Was this a translation of Dupree Lancome?

[00:18:00]Was there a French speaker? Yes, there were there was and his exemplars matched several other language patterns and the threat. as well. And we were in a big hurry on this case Because we wanted to make sure no one got hurt and this was a very very valuable pointer to a specific subject Very quickly.

[00:18:24]Roger did demographic profiling for the FBI when the Unabomber's Manifesto Came to light and An old sketch of the Unabomber was One of the many thoughts that they had over this endless investigation was that perhaps it was a younger person somebody who looked like this was at one of the bombing scenes and of course he doesn't look anything like Kaczynski because Kaczynski wasn't there.

[00:18:58]So Roger said the Unabomber is not as many thought a poorly educated young man, but was actually in his 50s highly educated with an advanced degree and the degree was not in humanities. The Unabomber, Roger continued, had lived in but he was not from the American West and he had been raised in Chicago in the American Midwest and he'd been raised a Catholic. And this he learned from reading the way the Unabomber used language in the manifesto.

[00:19:41]The Unabomber seemed to misspell words, but they weren't really misspelled.

[00:19:47]He was matching spelling reforms which was attempted by the newspaper Chicago Tribune during the 40s and 50s and Roger reasoned that a literate and intelligent Chicago area school boy coming up at that point might well adopt some of these spellings on his own as His own and that was one good reason that he was from Chicago and in his 50s, and that was true.

[00:20:11]He also said rearing children instead of raising children, which at that point was a northern cities feature. So several things pointed to Chicago and indeed he was from there.

[00:20:23]Also, he used Sierra as a topographical term. He said he went out to the Sierra into the Sierra or he went up to meditate or something like that, but he didn't use any other Western US terms such as ranch, fork, range, mesa and Roger reasoned that he was therefore not a native Westerner, but he was one who had moved to the west and indeed he had been and here is the rearing children as I said.

[00:20:55]Also, the Unabomber used 1960s terms playing footsie, working stiff, etc. He used sociological terms that said that he had a very good very intensive probably undergraduate education David Riesman's terms and on like that and Roger was right on all counts that he got from the linguistic evidence.

[00:21:26]Here's another case of both profiling at first and then authorship. This is the case of Charlene Hummert.

[00:21:39]Now, we've been talking about profiling.

[00:21:41]What can we tell about the person who might have written something and then we also have authorship at the end of that.

[00:21:49]So profiling and authorship are on this continuum.

[00:21:54]We keep narrowing down the suspect pool and then authorship analysis says who was the more likely person to ever written this. Now, this is the worksheet that we sent to you folks.

[00:22:18]And Remember our forensic linguistic investigative features and also remember and to beware of disinformation.

[00:22:33]So I'm gonna put this up on the video, but also if you have it there's a good time to open up that handout.

[00:22:46]So this is the first letter that was found on her husband's windshield Then considerably later she was murdered and as the investigation heated up this letter was sent to the police and the crass.

[00:23:10]So take a look now at what we said and I'm gonna stop share so I can read some of your questions.

[00:24:32]So these two letters seem very different.

[00:24:44]There is a ton of information in this letter, it's very wordy.

[00:24:51]There is no overt mistakes. I mean there's a couple of typos, but whereas this is clearly scrawled reminiscent of the Ransom note that we just saw in the prior case And The police asked me as they had asked Roger in the Kidnapping case. What can you tell us about whoever wrote these documents?

[00:25:37]So Since it took Benji Wald and me a few weeks I'll give you guys another five minutes or so.

[00:25:50]Of course even though I didn't know anything really about the case at that point I suspected that the police might suspect that they were written by the same person, but look how different they are.

[00:26:20]So after reading this around 9 million times, I Realized there was a connection. I mean there were obvious connections. It's written in English. Okay.

[00:26:33]There's no misspellings.

[00:26:37]Clearly there are funny things going on with this. For example the capital F for no good reason, but even then I knew, and this was one of my first cases, that that's the kind of easy thing that people might change to try to have people not think it's their writings.

[00:26:58]So I call this masking as opposed to masquerading which is when you make believe you're somebody else which happens, of course very often as well.

[00:27:18]So here we have a couple of dumbing-downs. "Cops have no idea how easy it is to pin husband when they only look there."

[00:27:28]But he uses determiners and prepositions quite well earlier on in the letter.

[00:27:49]Now people are pointing out very good things.

[00:27:54]For example, are there compound sentences are there non-compound sentences. But again are the object of this is to get a judge to allow the police to use other written evidence to make their case, and you need a search warrant for that and you need a good cause for that.

[00:28:20]So finally I noticed not only oh, I beg your pardon. I forgot this second document.

[00:28:35]If we want to assume that it is written by the first person or even not, is what the FBI Calls a POMIC. Jim Fitzgerald invented the term and it stands for post offense manipulation of investigation communication, otherwise known as a red herring, communication intended to divert suspicion.

[00:29:07]I killed her. I killed her not her husband, which is pretty good evidence according to several Homicide detectives when I would use this in training they'd say well gee I wonder who did it then. All right So here we have this again but now we have a lot of similarities and we also have what we call ironic repetition. A couple of days, I would have loved to have found out a couple of days later. She made sure my fiance found out.

[00:29:35]She wanted to break it off, so I broke her neck.

[00:29:41]This was obviously a rhetorical device and it turned out to be a rare rhetorical device See the rarer something is the more idiosyncratic it is, the harder it is to explain the way, if you find it in two supposedly unrelated documents, except as coming from a single author.

[00:29:59]And that's the object of authorship analysis: how likely is it that two or however many documents share and author have common authorship? And I went through Greek professors, Latin professors, all the way to the BYU encyclopedia of rhetoric and They said it was rare enough so that they had no specific label for it.

[00:30:21]So the judge granted us the search warrant and we got other writings from the chief suspect who unfortunately was as it so often is the husband of the murdered woman.

[00:30:39]When we got all that we notice Benjy noticed and took us a long time to notice this there was a total skewing of Contractions in all of hummert's workplace and prison writings and in the two anonymous letters which we call the Q Negatives were sometimes contracted but positives were never contracted So did not sometimes became didn't but I am was always I am never I'm. I will was always I will instead of I'll and These are some of the slides that I walked the jury through But remember there were many other Similarities there were the fact that both of these letters had very very skillful narrative manipulation and was so skillful flash forwards, flash backs, stepping aside that you didn't even know what was happening and that bespeaks a real skill and that skill was in both of those letters and Now we find this pattern that not only links the letters but it links to hummert's writings. I am it is I will we are, but don't. And we looked at Google, although I wouldn't use Google for Relative frequency anymore. But this is one of the first cases that I was using any corpus But we see here in the known writings. We have 74 non-contracted Verbs and 0 contracted and in the questioned documents the two letters, we have 23 and 0 positive contractions. now that's quite a skewing, quite an idiosyncrasy and we wondered what about in other, by other users, so the mass of Google, of course, we don't find it and interestingly in Google Scholar which is of course scholarly Papers I discovered to my surprise that there are almost no negative contractions. Look at the tiny number there in The negative contractions in Google Scholar.

[00:33:10]So to make a long story short It was not that everybody does that and I just never noticed it before or even that everybody in York, Pennsylvania did it and I never noticed it before.

[00:33:23]Now these analyses Compare competing hypotheses. That's how I restored at Columbia science progresses on.

[00:33:35]We try to explain the non-random distribution of the data by constructing competing hypotheses and testing Examining to see which better explains that non randomness.

[00:33:52]But forensic linguistics cannot say John Smith wrote something It just keeps shrinking the suspect pool. This is actually quite analogous to what the people of the Human Genome Project Tell me that DNA evidence does as well. They keep excluding people so this is a very important thing to realize. I Certainly, never identify a particular person as having read something I've written something but I do and others of course do we look at the Patterns and the cue and the patterns in the K And Hummert was, not just on the linguistic evidence, Convicted of Murder and sentenced to Life without parole.

[00:34:50]Several years ago. I got a call from a prisoner in the Chicago area who said that he had been convicted Largely on the basis of linguistic evidence and he'd been directed to me as possibly someone who could examine this evidence So my interns and I, and now this is grown to be the capital case death penalty Innocence Project at Hofstra, we have now done a number of these cases, but I'll tell you about the QB case in a moment.

[00:35:36]So my recent graduate and a PhD candidate at Aston Julianne Ford teaches the internship with me and we work with our constitutional law professor Eric Friedman and when necessary his interns and in analyzing evidence and appeal possibilities in capital cases where language evidence, typically recorded conversation, Interrogation or in the case what I'm about to tell you about a confession played a crucial role.

[00:36:08]We also work on other serious crimes for both defense and Investigators sometimes cold case murders we've done several of and recently we helped track and charged a serial killer in the American South.

[00:36:26]Someone was killing her husbands Making it look like a suicide According to the police and we were able after many months study of everyone's writings to Conclude that the superior hypothesis. Was that she had or I should say that the Patterns in her writings were the best explanation for the patterns in one of the husband's suicide notes.

[00:36:57]So now Antwaun Cubie is the man that got me involved in this.

[00:37:05]So 1996 accused of murder and armed robbery From various pieces of evidence the most important one and the only one that has not Well, it's a long story. But anyway here is the confession and Here is Cubie's signature The only problem is that none of the I shouldn't say none, but the patterns of language in this confession do not match Cubie's, but they do match the detective who testified against him.

[00:37:45]Cubie maintains he was interrogated and beaten and beaten and beaten and unfortunately Chicago police were Infamous during that period of time there were many court cases for torturing Defendants.

[00:38:02]Finally he was told okay you want a phone call, sign these forms.

[00:38:06]He signed the two pieces of paper and the next time he saw them there was confession typed over his signature so The authorship of the alleged confession is questioned.

[00:38:19]So my interns we didn't have the death penalty Innocence Project yet, but we were working on other cases and I analyzed So the competing hypotheses here were the patterns of the Q documents Better explained as being instances of which patterns of Cubie's known writings or other than that for example, the language of the police officer who testified against him and On various measures and I'll just zip through these We found these features in the confession do not appear in Cubie's known writings and Certain of them do appear in the speech patterns of the officer and in the general register or genre that other people Described as police writings.

[00:39:24]So complementizer deletion and for non linguist, I just won't go into it but linguists can see The pattern matched the did not match Cubie Contraction patterns the patterns did not match Cubie even though the We would expect more contractions in Verbal rather than written and the police said that Cubie had dictated it to them, but we find the exact opposite pattern Discourse markers not a convincing match and then Comparing the confession to the language of the police officer approximately They have QB saying now it's very very common of course for police to take a Confession and help the person get exactly what they want to say And there's nothing wrong with that as long as there's no coercion and everything else but the police here maintained that was not what happened They did not say that any single word was written by them or suggested by them They said they testified each and every word that Cubie spoke was what was in that confession.

[00:40:48]Yet we find approximately and add an unknown time And then we also have an interesting syntactic structure Which is the placement of then typically we'll say something like we ate dinner then we went home We typically don't say we ate dinner we then went home we could But it's not canonical. It's not the most normal in in most sentences and What we find and I'll summarize for you is that Cubie has the more normal.

[00:41:18]Then he shut up then I said then the little guy inside but the officer has the all three then went he then walked, and what do we find in the confession? I then told we both then went.

[00:41:35]So what is the possibility of this eighteen-year-old guy Because he's giving a confession to murder all of a sudden Using a syntactic structure that we have no evidence that he ever used before matter of fact, he barely uses the term then.

[00:41:56]So we see of course an absence of the number of matching links between the Q and Cubie's K and a presence Between the Q and the K of the testifying author. So this supports Cubie's contention. He was not the author of The confession that brought him a life sentence but he had exhausted all his normal appeals.

[00:42:16]Then the state of Illinois passed a law that said if you are tortured during that period of time you had more appeals and this is a 2016 letter and they were very interested in the analysis of the confession as well. And Cubie called me last week and said that things are going fairly well and on track so now he has been in jail since 1996 and A good six years ago, we showed that How highly unlikely it was that he had written his own confession.

[00:42:51]Okay now here is a phrase of my partner in not crime Tammy Gales Fabulous corpus linguist and forensic linguist and theoretical linguist who runs the program here with me.

[00:43:04]No one knows they need a linguist, but everyone needs a linguist and I thought it might be interesting Especially cuz I'm so proud of my graduates. Where have Hofstra forensic linguistics Graduate gotten jobs? huh and somebody forgot to spell linguistics correctly. Well, that was me.

[00:43:20]Okay private intelligence agencies, law firms, teaching, freelance, FBI, CIA, NCIS, you name it and students also go on to PhD programs in the US and abroad.

[00:43:40]I'm gonna end with a bomb threat case that shows sometimes patterns... You're looking for patterns and you're looking in the wrong places.

[00:43:53]Schemas, well linguist knows that schemas are narratives or or or Knowledge structures through which we understand the way the world works can be limiting. So in Beverly Hills a house started to receive bomb threats.

[00:44:15]Here's one of them. We'll see it again Here's another one Here's another one. Ah Jodie Foster said I now many people might remember that Jodie Foster, a man tried to kill President Reagan because he said I'm in love with Jodie Foster an actor and she doesn't know I'm alive and now she'll know I'm alive, so Here are the three bomb threats. now without going into the way the *inaudible* analyzes threatening communications We don't have personalization here, but we have repeated threats Well, anyway, I'll just stick to This very case. So the owner of this house A very rich person wasn't even there His head of security, you guys may or may not know that many very rich persons have their own police forces for good reason and They said well we are out of our Element here. We just don't know what to make of this so they went to another group of private intelligence agency composed of ex-CIA agents and somebody had obscenely speedy, so They asked me if I would help and I said, of course so first thing I asked by the way in all cases like this is have you told the police because I want you to tell the police we have to have everybody possible involved and they said yes they had. Okay, so I'm driving back and forth to Hofstra and they're telling me what's on these notes and then another note comes another day and they read it to me and they say look at this capital B in bomb. Yes, okay, but I mean, it's just like with the hummert letter.

[00:46:18]So we saw maybe a certain kind of set instructor, but what do we do then? Who do we say wrote this I mean there's too many people in the world who might Capitalize both the B's and bomb and we don't know who they are.

[00:46:32]Anyway So what you try to do in a case like this is you try to get something that is outside the four corners of the letter in the real world, and that was why I was Very interested in the fact that we had Jodie Foster So what else in here is Linked to something in the real world Okay. Well this house clearly okay, but that's the house that's receiving the threats.

[00:47:04]Gas bomb perhaps So I said hmmm gas bomb... only gas bomb incident I know off the top of my head is the sarin subway attack in Tokyo. I said look, have the police look at the BAU's CTAD, the communicated threat assessment database which the Behavioral Analysis Unit maintains Jim Fitzgerald Began it and I worked on it.

[00:47:34]Go through the police to the local FBI field office you have to do it through channels and they will request their BAU Look in their over 4,000 threats and other criminally oriented communications for the term gas bomb. We may find that John Smith just got out of jail in Pittsburgh and he had been put into jail for gas bombs or at least Threatening with gas bombs and they said sorry we can't do that. I said, what are you talking about?

[00:48:01]We're talking about a bombing here. What do you mean you can't do it and they said our police don't talk to the FBI.

[00:48:07]I said what? What is this some bad Late-night TV show where the local cops are are against the FBI This could be the clue that you want. You keep on telling me. What do we have in these letters and that's what we have.

[00:48:21]Now we also have other things right as I'm sure you've noticed the Writing goes up and down instead of sideways except for Jodie Foster and then we have some other odd things. Now I noticed things just like I did with the narrative because of the language is that I've studied and my Specialty languages Swahili in East African Bantu language. I took Swahili in college because out of 65 or 55 and remembering more languages toward a Columbia was the only one wasn't taught five days a week and I was in a Performing group and I couldn't have classes Monday or Friday and yet that was what happened in college those days so there was one that wasn't taught Monday through Friday and that was Swahili and as I love telling my Swahili students I've taught it now in university level for decades, the day I walked into Swahili class the first day I could not have even found Africa on a map That's how much I knew about Swahili or Africa but I fell madly in love with it has 15 genders the what would be a gender if it was a Romance language Obviously not having to do with sexual gender, but neither is it in La Mesa But *French* is a table feminine in Spanish and masculine in French. It's just a way of keeping track of things so three of those Genders have to do with different aspects of space and time and Swahili. I mean, it's a linguist dream So I fell madly in love with it. I wound up taking it Specializing it in graduate school then I got a Fulbright fellowship for one year to Kenya and I stayed there seven so one of the things that I wrote about was Swahili demonstratives. The other thing I wrote about was Swahili narratives which really helped me in the hummert case because Swahili is very interesting different set of Tenses that you use when you have narrative sequences So I see those things and that's why I immediately saw in hummert that the narrative sequence was being handled so brilliantly and here I had to pay a lot of attention to demonstratives. So here we have this house this house this house We also have here this house what?

[00:50:39]so I kept wondering What is this? What am I missing? There must be something that is being communicated here that I'm just not getting.

[00:50:53]Finally I said perhaps it's because I'm making an assumption. I am viewing these as letters what if they're not letters what if they're poetry what if they're something else and On a whim I looked up Well before I get to that as I said one time. I talked to the Yale Sherlock Holmes Society. Imagine how much fun that was is great and Sherlock will always say things like this. He'll look at these. He says aha. Yes there's to be a bomb but not at this house not of the house that's receiving the threats, but at Studio City shopping area on Ventura Boulevard Clearly right. How do we know that? Well, I Said what if this isn't a letter, but if it's a map Now here is a jodie foster's house which I did not get from my intelligence friends, but I found on the way on a website and here anonymized a little bit is The house that was receiving the threats Then I took the Jodie Foster threat and Superimposed it on this and there we are Jodie Foster over at a this house down by B and a gas bomb at Studio City and No one could ever have been more surprised than I when that worked. and I sized it but of course I didn't move anything And in terms of relative space So I immediately called these guys up and at that very moment The letter writer was recognized by somebody's chief of security as a previous stalker And he said oh that's so-and-so We knew him from this other city. This is exactly the kind of letter that he sent or Something and he was picked up and he never talked. He never told anybody what the letters meant why they were a map or were not a map never gave any information at all and Luckily, no bomb went off at Studio City So to recap we try to help non linguists gain maximum intelligence From language evidence and If people are interested in addition to all our BA MA, MA, MA JD we also offer intensive one-week Courses which when Jim Fitzgerald retired from the FBI we moved his week-long boot camps that he had invented and developed to Hofstra and in The spring he teaches one with me and very heavily oriented on threat assessment and FBI procedure and tanya christiansen from the university of copenhagen Who with whom I've worked in many cases now She and I teach the one in The fall which is more internationally oriented in European cases and is also more generally Introductory to forensic linguistics.

[00:54:42]So I hope this is a bit of interest to you all I'm gonna look at chats now, and I thank you for attending.

[00:55:02]People can apply for the masters still beginning in the fall we saw a couple of spots I Don't know about online courses it's a very complicated question because the quote-unquote distance learning courses must adhere to strict formats But if you are interested please email forensic linguistics at Hofstra.edu and my unbelievably good graduate fellows and our admissions people and I will answer any other questions about courses and like that.

[00:55:52]People are very kind.

[00:56:11]So, yes, you can find Jodie Foster on Google as I say I like to say, you know I called up British intelligence and they told me where she lived but yeah, you can find you know, where all the stars houses are So, I'm very happy that people found this interesting.

[00:56:28]Yes will be a bomb versus going to a bomb notice that - of course, but again It didn't help us know who was sending the the bomb.

[00:56:44]This is great and I see a lot of old friends here, I'm happy to see you all and I'm glad you all seem to Be interested and if you have cases that you think we might be interested in helping with Please let us know as I said forensic linguistics at hofstra dot edu. And The this presentation will be cleaned up and edited and posted on the Hofstra YouTube channel I believe within a few days.

[00:57:20]Why is there a difference between the three bomb letters? Ah you got me.

[00:57:26]That was a very puzzling set. It did of course look like they were all sent from the same person And Talking about how many international students are accepted for the master's studies. That's all admissions knows all the details of that we have no restrictions ourselves as The person asking me that maybe there are travel restrictions.

[00:57:56]And of course, I would need a crystal ball to know just what's gonna happen from there.

[00:58:08]Hi Hannah.

[00:58:11]Hi Sandy, it's very good. I see a lot of old friends here. It's very very nice. I Have done a lot of work on nine-one-one and I was on a Nightline talking about it a little girl disappeared in Colorado and the Father Was giggling when he was reporting her missing and And he said oh and I called her mom and I told her to get her butt over here ha ha ha ha and little girl wasn't found. A screen had been taken down from her room and as you can imagine everybody really piled on to the dad because he didn't fit their expectation of What a father should sound like with 911. I Said you cannot tell how people are gonna react to stress and they had an FBI agent on who also said the exact same thing and as a matter of fact One of the Cases that I used to illustrate it Was I'd have the book here, but I can't put my hand on it now in this moment Was Marty Tankleff who at age 17 came downstairs First day of senior class in high school to find his parents bludgeoned to death in pools of blood And he called up 911 and it was used against him in the trial in which he was falsely convicted Spent 17 or 18 years in jail until he was fully exonerated Because he was too calm and in his 911 absolutely cannot tell from that. And also there's a new show called Reasonable doubt Which looks at Looks at old cases and they had me on to talk about a 9-1-1 case So that should come out whenever in the reasonable doubt does again Yeah, there's a lot of work that needs to be done on 9-1-1 calls. Mm-hmm Yeah sandy. Sandy, doesn't it right. So thanks for doing this Wish we had the publicist that other corners of forensic science. two Lawyers have been heard to say I wish I knew about forensic linguistics X years ago. Yes That's exactly why time he says nobody needs a linguist, but everybody needs a linguist Yeah, so people ask very reasonable questions about do we use software yes To cut to the chase we are trying to develop software that does what Tania and Jim and Benjy and Tammy and Dakota Wang and Julianne Ford and I do and There's a lot of really good software out there computational linguists habit, but it's a long story but So far, there's a mismatch. Obviously we can't have programs do exactly what we do because there are so many variables But that's why we're developing the Flint engine which we are trying to Use AI to Encode that which we actually do in other words, not nearly by AI black boxes and things nothing against AI black boxes are supposed but That that's a whole other conversation For ineffective assistance of counsel how long has this information been known? Well Roger shy wrote a book in 97 Which is a seminal book and in it, he recounted many cases that he had already worked on. So If it had occurred to somebody to look around for a forensic linguist the Expertise would have been there for quite a while now Yes, Jenny, it's a good point One of my colleagues at Columbia at Hofstra writes I don't believe a computer would have come up with the idea of a map But that's a really really good example of what computers don't do they can't take off you have to tell a computer what to do and When you realize how many?

[01:03:33]infinities of variables we're always taking into account and also how little information is Encoded in words much of it is in our own minds and we read into it. And again, that's a Schema and greicy and Maxim's. That's a whole other conversation Yah, and we've worked with a lot of mitigation specialist death-penalty MIT specialists And if you have cases Please reach out to Eric Friedman the professor of constitutional law and he is The conduit and the fact the person who can vet cases and and because he understands what postures of cases mean I mean Unfortunately if I had known him To work with I knew him, but when I got the Hubie case I probably would have told QB that there was nothing to do because he had supposedly used up all his His appeals and he had but then the law changed Yes, I worked on the later part of the JonBenet Ramsey case and that's again a very very long story I Would refer people to Jim Fitzgerald's opinions on the JonBenet Ramsey case, what a What a sad case Huh so Andy Disney says I talked about Rajesh I hid napping case in class and a student informed me It had been on CSI the week before but they didn't give Raj or any royalty. So I'll let him know. That's pretty good Of course, he did publish it in Dialect I Work on speaker identification. I'm not a phone to Titian if you're interested in a fabulous politician who works on speaker identification of look up Sandra dinner at USC however, I've worked on a number of cases because the politicians and audio engineers that I've worked with in the past Knew that it was useful to have somebody who could be able to describe dialectal variation last case I worked on like that the Two people who were accused of being the same person spoke such different dialects demonstrably that It was very hard to believe they were the same person Again, that's a whole other conversation Book and reading Recommendations take a look at our website and our FAQ Because there's a whole bunch of really excellent books. Most of them Rajesh. I Rodgers written about thirty books now on many cases The language of bribery cases perjury cases of defamation cases Language of murder cases which he dedicated to me. So I'm very happy for that Malcolm Cole card Johnson's book is excellent the Speaking of crime he's going to come out in a new edition soon and that will be great too Very good cuz I'm I'm going horse So I will sign off now But I so appreciate all these excellent comments and questions and I wish I could go through them all Individually and at great length and maybe some at some future point. We will be able to Okay, I'm gonna Leave now bye everybody. Thanks for coming